Antibiotic Lock Technique: Review of the Literature

Megan B. Bestul, PharmD; Heather L. VandenBussche, PharmD


Pharmacotherapy. 2005;25(2):211-227. 

In This Article

Antibiotic Lock Technique

The concept of antibiotic lock technique was developed in the late 1980s as a derivation of the heparin lock technique. This method involves instilling a highly concentrated antibiotic solution into a catheter lumen and allowing the solution to dwell for a specified time period for the purpose of sterilizing the lumen. One group of researchers hypothesized that using high concentrations of bactericidal antibiotics with activity against common catheter-related bacteria could effectively sterilize a catheter and reduce catheter-related sepsis from intraluminal line infections.[45,46] They determined that the vancomycin and amikacin concentrations used in the first antibiotic lock technique experiments were 40-80 times and 60-120 times greater, respectively, than the peak blood concentrations attained when the antibiotics were administered systemically with conventional dosing. Other studies documented that concentrations of vancomycin and amikacin in a catheter lumen maintained high levels for at least 8-12 hours and were stable and microbiologically active over a 12-hour dwell time within a lumen.[46,47]

Antimicrobial choices for use in the antibiotic lock technique are dependent on the different pathogens suspected to infect the catheter lumen ( Table 1 ), characteristics of the organisms (i.e., ability to produce slime, adherence to host proteins), and the pharmacodynamic properties of the antimicrobial agent. For antibiotics such as aminoglycosides and fluoroquinolones, known as concentration-dependent killers, optimal and rapid bacterial killing occurs at high concen-trations (peak concentrations of at least 8-10 times the minimum inhibitory concentration [MIC] of a targeted organism or, for fluoro-quinolones, a 24-hour area under the concen-tration-time curve:MIC ratio of ≥ 125 for gram-negative organisms).[48] As a result of the postantibiotic effect of these drugs, maintaining high concentrations throughout a dosing interval or dwell time within a catheter may not be necessary. In comparison, b-lactams and vancomycin require maintenance of a concen-tration above the targeted organism's MIC for most of the dosing interval for optimal killing and are, therefore, known as time-dependent killers. Ideally, antibiotics used in antibiotic lock technique may be most effective if high drug concentrations are maintained for an extended period of time to maximize killing ability since high antibiotic concentrations may be needed in order to penetrate the microbial biofilm.[3] As a result of these stipulations, stability and compatibility become important factors to consider when selecting an antibiotic. The antibiotic(s) must be stable within the catheter lumen for the dosing interval or dwell time, as well as be compatible with the type of catheter used and any other component present in the lumen, such as heparin or other antimicrobials.

Secondary to the presence of certain host proteins in the catheter lumen, such as fibronectin, fibrinogen, and fibrin, heparin may increase the efficacy of antibiotics used in antibiotic lock technique to treat catheter-related infections.[36] However, the concept of combining heparin with antibiotics in lock solutions is as controversial as the use of the antibiotic lock technique. Physical compatibility and chemical stability of the components of the antibiotic lock solution, as well as heparin use in certain patient populations (i.e., hypocoagulable states, documented heparin-induced thrombocytopenia), are limiting factors that need to be considered when adding heparin to an antibiotic lock solution. Numerous studies have been conducted that include heparin as a component of antibiotic lock technique solutions.[49,50,51,52,53,54,55] One group of investigators who are considered pioneers in the development of the antibiotic lock technique in the 1980s primarily studied heparinized antibiotic lock solutions, whereas another group conducted research by using nonheparinized antibiotic lock therapy.[45,46,50,52,56,57,58,59] Subsequent studies have cited the original work regarding specific antibiotics and study methodology for conducting antibiotic lock research.[47,49,51,53,55,58,60,61,62,63,64,65]

Numerous in vitro studies have evaluated different concentrations of various antibiotics and their activity against common bacteria that infect catheter lumina ( Table 2 ).[49,50,52,53,60,65] Data from these studies have provided information regarding effective antibiotics, dosages and concentrations, and dwell times that have been used in clinical trials.

One group examined the stability of a vancomycin and heparin solution and its antibacterial activity against staphylococcal isolates.[50] Vancomycin stability over time was measured directly by fluorescent polarization immunoassay and indirectly by determining the MIC and minimum bactericidal concentration (MBC) against S. aureus and Staphylococcus epidermidis on days 1, 12, 26, and 40. On day 85, MICs were measured again for all solutions. In addition, the authors studied heparin stability in the vancomycin-heparin and control solutions by measuring the activated partial thromboplastin time (aPTT) at 0, 3, and 6 weeks. The vancomycin concentrations and the S. aureus MICs were similar for all solutions throughout the study period. However, the vancomycin-heparin solution was not bactericidal against S. aureus at vancomycin concentrations of 12.5 µg/ml, whereas the control solution MBCs were between 3.12 and 12.5 µg/ml. Against S. epidermidis, all vancomycin-heparin and control solutions were bactericidal at concentrations of 1.56-6.25 µg/ml. Prolongation of the aPTT was similar at all time periods and was maintained up to 5.5 months. The authors concluded that the vancomycin-heparin solution maintained antistaphylococcal and anticoagulant activity for at least 85 days whether stored at 4°C or room temperature. This study is limited by its lack of evaluation using a true model of catheter infection.

Another group reported the effectiveness of vancomycin antibiotic lock therapy versus conventional systemic vancomycin therapy in treating polyvinylchloride catheter surfaces that were colonized with S. epidermidis.[65] After induced colonization, a parenteral nutrition solution was infused through each catheter at room temperature for 9 hours/day for 3 days. After each infusion, the catheters were clamped and incubated at 37°C for 15 hours. Catheters were then randomly treated with vancomycin alone or in combination with netilmicin, fosfomycin, or rifampin for 3 days, whereas control catheters were infused with only nutrition solution (five of each type). Vancomycin was administered by either intermittent administration (1-hr infusion alone or sequentially in combination with another antibiotic every 8 hrs with continuous nutrition solution infusion at Y-site), by 24-hour continuous administration (vancomycin added to the nutrition solution), or by antibiotic lock technique (2.5 mg/0.5 ml injected twice/day and locked until the next antibiotic administration). After the 3-day antibiotic therapy, each catheter lumen was rinsed with 50 ml of parenteral nutrition solution, clamped, and incubated at 37°C for 15 hours to allow bacterial replication before colony counts. Catheters were flushed with tryptic soy broth that was placed on tryptic soy agar plates and incubated for 24-48 hours. A catheter was considered sterile if the broth did not contain any colony-forming units (cfu).

When intermittently infused, vancomycin had a statistically significant effect on bacterial growth (p<0.001) but did not sterilize any of the catheters. Vancomycin in combination with other antibiotics reduced bacterial growth compared with vancomycin alone (p<0.05), but sterilization of the catheter lumina was inconsistent within each combination and was variable among antibiotic combinations. Continuous infusion of vancomycin also reduced bacterial growth but failed to sterilize the catheters. Vancomycin antibiotic lock therapy sterilized all five of the catheter lumina but did not produce a statistically significant decrease in bacterial growth compared with vancomycin administration in combination with the other antibiotics. In a separate time-kill experiment, high concentrations of vancomycin (450 or 5000 mg/L) resulted in rapid bactericidal activity versus a slower kill rate with a lower concentration (7.5 mg/L) alone and in combination. The authors concluded that high antibiotic concentrations in close contact with pathogens for extended periods of time provide more effective bactericidal treatment than that of lower concentrations in solutions that flow over bacteria during intermittent or continuous infusions. They stated that systemic antibiotics were required if there was evidence of extraluminal infection since antibiotics do not significantly enter the bloodstream when using the antibiotic lock technique. This study compared antibiotic locks with conventional modes of therapy by using an appropriate catheter infection model but is limited by its small sample and evaluation of only S. epidermidis colonization.

Researchers also evaluated a vancomycin-heparin-ciprofloxacin solution for its stability and activity against gram-negative and gram-positive organisms.[52] Preparation of the vancomycin-heparin portion was identical to that of a previous study[50] with the exception of a small decrease in the heparin concentration (9.75 vs 9.73 U/mL). In addition, ciprofloxacin 2 µg/ml was added. Vancomycin concentrations were measured directly by fluorescent polarization immunoassay, and aPTT was measured to determine indirectly the stability of heparin. In addition, the MICs and MBCs of vancomycin (for S. aureus and S. epidermidis ) and ciprofloxacin (for Escherichia coli, Klebsiella pneumoniae, and Pseudomonas aeruginosa ) were determined to evaluate indirectly the stability of the antibiotics in the various preparations.

The combination solutions were as effective as vancomycin or ciprofloxacin alone in inhibiting or killing all bacteria tested. Doubling the vancomycin concentration to 50 µg/ml did not result in a lower MIC for either S. aureus or S. epidermidis. Ciprofloxacin did not appear to have a synergistic or additive effect with vancomycin against gram-positive bacteria. For gram-negative bacteria, the ciprofloxacin activity in all antibiotic lock combinations was identical to the activity of the ciprofloxacin control. The addition of ciprofloxacin and/or vancomycin to heparin did not affect anticoagulant activity when compared with the standard heparin lock solution. The solutions maintained anticoagulant activity for at least 3 months and were not affected by storage temperature. The researchers concluded that both antibiotics maintained full antimicrobial activity and were appropriate concentrations for prophylaxis of catheter-related sepsis when combined with heparin. This study is limited by its lack of evaluation using a catheter infection model.

Another group examined the ability of various antibiotics to reduce or eradicate bacterial colonization in a silicone catheter lumen.[60] Strains of S. aureus, S. epidermidis (three different strains, one that produced a highly adherent biofilm), K. pneumoniae, Enterobacter aerogenes, C. albicans, and Candida tropicalis were used to infect the catheters. The catheters were filled with total parenteral nutrition solution and incubated for 12 hours at 35°C followed by instillation of the antibiotic lock technique for 12 hours and repeated for 7 days. The antimicrobials used are shown in Table 2 . Catheter microbial adherence was assessed over 12 hours, and catheter segments were analyzed for quantitative microbial counts on days 1, 4, and 7.

In a separate experiment, the antibiotic lock solutions were bioassayed after 4, 8, and 12 hours of incubation to determine the intraluminal antimicrobial concentrations over time. Each antiinfective was stable, retaining greater than 90% activity over the 12-hour incubation period. All test strains were sensitive to the appropriate antibiotics. Microscopy revealed that E. aerogenes was less adherent to the catheters than were S. epidermidis and K. pneumoniae, which produced biofilm. After an antibiotic lock with gentamicin, nafcillin, ceftriaxone, or vancomycin, a statistically significant decrease was noted in S. aureus and S. epidermidis (the biofilm-producing strain and one strain without biofilm) colonization or elimination by day 1 (p<0.001). Colonization with the third S. epidermidis strain (a nonslime producer) was reduced by day 1 when treated with nafcillin or vancomycin and by day 7 when treated with ceftriaxone or gentamicin. The K. pneumoniae colonization was eradicated by day 1 with aztreonam, ceftriaxone, and gentamicin (p<0.001). Aztreonam and ceftriaxone reduced E. aerogenes colonization by day 1 as compared with day 4 with gentamicin. Both yeast strains were successfully eradicated by day 7 with the use of amphotericin B, whereas fluconazole eradicated C. albicans and significantly reduced C. tropicalis colonization by day 7. The researchers concluded that the reduction in microorganism colonization was statistically significant for gram-positive and gram-negative bacteria and yeasts after 7 days of appropriate antibiotic lock therapy. Although the study involved a catheter infection model using multiple antimicrobials and microorganisms, it used only high drug concentrations without heparin in the study solutions and assessed stability by using bioassay rather than chemical assay.

In 1999, another group reported the stability of vancomycin, cefazolin, ticarcillin-clavulanic acid, ceftazidime, or ciprofloxacin combined with heparin in test tubes for use in port infections ( Table 2 ).[53] All antibiotic concentrations were 500 µg/ml, except ciprofloxacin owing to macroscopic precipitation observed at higher concentrations. Heparin stability was not assessed in this study. In separate test tubes, 104 cfu/ml of susceptible bacteria ( S. epidermidis, K. pneumoniae, E. coli, or P. aeruginosa ) were added to each combination to assess if stability was altered in the presence of microorganisms. Antibiotic concentrations were determined by bioassay on days 1, 3, 7, and 10. After 1 day, antibiotic activity remained constant for all solutions, regardless of bacterial presence. There was a 28-36% decline in ceftazidime activity after 7 days and up to 50% loss of activity after 10 days for samples stored at 37°C. There was 10% or less reduction in antibiotic activity of all other agents at both temperatures. The authors concluded that the antibiotic solutions were stable since the concentrations remaining after 10 days were significantly above the MICs of common line pathogens, including that of ceftazidime which may be adequate for dwell times of up to 7 days. However, this study did not evaluate chemical drug stability or heparin activity over time and did not use a true port infection model.

Another group assessed the stability of cefazolin, vancomycin, ceftazidime, ciprofloxacin, and gentamicin for use in an antibiotic-heparin lock solution ( Table 2 ).[49] Each antibiotic was stored alone in separate glass test tubes for 72 hours and was analyzed at 24-hour intervals for stability by using spectrophotometry. Antibiotic-heparin solutions were then prepared and stored in both glass test tubes and dual-lumen polyurethane CVCs for 72 hours, analyzed for stability with spectrophotometry, and confirmed with high-performance liquid chromatography.

For antibiotics stored alone, absorbance was reduced for cefazolin (6.7%, p<0.05), ceftazidime (13.4%, p<0.05), gentamicin (3%, p<0.05), and vancomycin (0.7%, NS). After heparin was added, cefazolin was the only agent with an additional significant decrease in absorbance of 2.4% (p<0.05). Ciprofloxacin was removed from the study because a precipitate formed immediately with the addition of heparin. The most significant absorbance reductions were seen with combinations stored in CVCs: 27.4% with cefazolin, 29.7% for vancomycin, 40.2% for ceftazidime, and 8% for gentamicin (p<0.001), suggesting adsorption to the catheter surface. The authors stated that these reductions may not be clinically significant because the concentrations remaining in the catheter lumen were approxi-mately 5 mg/ml (substantially greater than the usual MICs of organisms involved in causing catheter-related infections). They concluded that high antibiotic and heparin concentrations were stable inside CVCs over 72 hours. This study is limited by its use of only high drug concen-trations, no evaluation of heparin activity, and its lack of assessment of antimicrobial activity against microorganisms and effects with use of a true catheter infection model.

These in vitro trials provide information regarding the stability of antibiotics when used alone, in combination with other antibiotics and/or heparin, and in different catheter devices for antibiotic lock treatment. In addition, antimicrobial efficacy was demonstrated in some studies by reduction of bacterial colonization and stable bactericidal activity over time. Most of the antibiotics evaluated maintained stability and/or efficacy in polyurethane CVCs for at least 7 days. However, ciprofloxacin stability in antibiotic lock solutions was dependent on its concentration, which must be considered in any therapy that includes its use.

These studies differed significantly in the antibiotic concentrations used, varying as much as 1000-fold, and did not always include an anticoagulant, which may be a desirable component of antibiotic lock therapy. In addition, the study models used variable drug stability assessment methods and catheter infection models. They also inconsistently evaluated the effects of bacteria (including biofilm producers) in the study models. Despite these differences, these studies have served as guides for antibiotic selection and concentrations used in human trials of the antibiotic lock technique.

The clinical studies can be separated into two categories based on the use of antibiotic lock technique for prevention or treatment of catheter infections ( Table 3 and Table 4 ).[45,47,51,54,55,57,58,59,61,62,63,64,66] In comparison to the laboratory studies, most of the treatment studies do not include heparin as a component of the antibiotic lock solutions.

Prevention Trials. One group performed a randomized, double-blind study in pediatric hematology or oncology patients to evaluate the effects of a vancomycin-heparin flush solution on the occurrence of bacteremia resulting from CVC colonization with vancomycin-susceptible bacteria ( Table 3 ).[58] Children were randomly assigned to receive either vancomycin-heparin or heparin alone for all catheter flushes. Episodes of fever or sepsis were evaluated to determine if systemic antibiotics were warranted. Children remained in the study until either the CVC was permanently removed or when study solutions could no longer be administered.

There were 45 children enrolled, with 52 catheters placed: 21 patients received vancomycin-heparin and 24 received heparin alone. Six episodes of bacteremia related to luminal colonization with vancomycin-susceptible bacteria were reported in the heparin group compared with no episodes in the vancomycin-heparin group (p=0.035). This was defined as a peripheral vein culture with 10% or less of the organisms obtained when drawn through the CVC and with no evidence of exit-site infection. Cultures from one patient in the vancomycin-heparin group grew K. pneumoniae, whereas all six cultures in the heparin group grew vancomycin-susceptible organisms (five coagulase-negative staphylococci and one Corynebacterium sp; in addition, one culture also grew E. coli ). There was a significantly longer time to the first episode of vancomycin-susceptible bacteremia in the vancomycin-heparin group compared with the heparin alone group (p=0.04). No vancomycin-resistant organisms were isolated during the study. Peripheral blood vancomycin levels were undetectable, but when the blood specimens were drawn in relation to flush solution administration is not clear. The authors concluded that use of vancomycin-heparin flush solution in immunocompromised children with indwelling CVCs may reduce the frequency of vancomycin-susceptible bacteremia related to catheter colonization. One limitation of this study is the variability in dwell time between patients since the use of catheters depended on individual patient needs.

In a double-blind, randomized trial, the effectiveness of antibiotic lock technique in preventing catheter lumen colonization and subsequent gram-positive bacteremia was evaluated in 117 adult patients with chemotherapy-induced neutropenia (neutrophil count < 500 cells/mm3).[51] Patients with nontunneled, multilumen, polyurethane CVCs were randomly assigned to receive heparin (57 patients) or vancomycin-heparin (60 patients) lock solutions for an average of 10-11 days. Exclusion criteria were patients with clinical or microbiologic evidence of infection, vancomycin allergy, or need for antibiotics or parenteral nutrition. Catheter hub and insertion site cultures were collected at baseline and twice/week and were repeated along with blood cultures before starting systemic antibiotics for those who developed febrile neutropenia. Catheter tip cultures were also performed for any removed catheters.

Significant catheter hub colonization (≥ 15 cfu/ml) occurred in 9 (15.8%) of the 57 patients in the heparin-only group compared with none in the vancomycin-heparin group (p=0.001). Colonizing organisms were S. epidermidis (7 patients), Staphylococcus capitis (1 patient), and Corynebacterium sp (1 patient). Catheter-related bacteremia occurred in 4 (7%) of 57 patients receiving heparin ( S. epidermidis in 3 patients and S. capitis in 1) compared with none of the patients in the vancomycin-heparin group (p=0.05). One patient with S. epidermidis colonization required catheter removal because of breakthrough bacteremia. None of the isolated organisms were vancomycin resistant. The proportions of patients who remained free of catheter hub colonization and catheter-related bacteremia at study end were 74.6% and 88%, respectively, in the heparin group compared with 100% for both in the vancomycin-heparin group (p=0.004 and 0.06, respectively).

The authors concluded that vancomycin-heparin antibiotic lock therapy is successful in decreasing the frequency of catheter hub colonization with gram-positive bacteria in neutropenic patients. They also concluded that the local administration of vancomycin during periods of chemotherapy-induced neutropenia decreases the risk of resistant bacterial growth compared with systemic antibiotic administration. However, it is important to note the short treatment duration used in this study, making the probability low for selecting or producing resistant organisms.

Another group reported on the efficacy of a vancomycin-heparin flush solution for prevention of CVC-related bacteremia in 83 oncology patients with a single-lumen CVC.[66] The patients were randomly assigned to receive a daily flush with vancomycin-heparin (39 patients) or heparin alone (44 patients). Febrile episodes were evaluated, and central and peripheral blood cultures were drawn before starting systemic antibiotics.

Sixty-four patients experienced 143 febrile episodes (82 episodes in the heparin-alone group and 61 in the vancomycin-heparin group). More episodes occurred in nonneutropenic patients receiving heparin alone (35 episodes) than in those receiving vancomycin-heparin (14 episodes, p=0.014). In neutropenic patients, there were 47 episodes in both groups. Bacteremia was documented in 44 cases in which 23 were gram-positive organisms, 20 were gram-negative organisms, and 1 was Candida. Bacteremia with vancomycin-sensitive organisms was reported in 16 episodes in the heparin-alone group compared with 7 episodes in the vancomycin-heparin group (p=0.19). No episodes were reported in the vancomycin-heparin group of nonneutropenic patients, whereas 9 of the 16 episodes in the heparin-alone group were in nonneutropenic patients (p=0.019). No vancomycin-resistant organisms were found during the study period.

The authors concluded that vancomycin-heparin flush solutions effectively prevented bacteremia with vancomycin-susceptible organisms in nonneutropenic patients but had limited value in patients with neutropenia. Limitations of this study include a small sample and lack of distinction between catheter-related bacteremia and bacteremia due to other causes, including contamination.

In another study, the authors evaluated the effects of a broad-spectrum antibiotic flush solution on prevention of central line infections in 126 pediatric oncology patients.[59] Children were excluded if they had totally implanted catheters, were critically ill, or had continuous fluids running through the CVC. Children were randomly assigned in a double-blind fashion to receive vancomycin-heparin-ciprofloxacin (34 patients with 38 lines), vancomycin-heparin (28 patients with 35 lines), or heparin alone (64 patients with 80 lines). Flushes were used for the entire study or the life of the CVC. Once a patient became febrile, each febrile episode was evaluated and treated according to criteria that included assessment of neutrophil count, presence of sepsis, and evidence of CVC infection.

The line infection rate (possible, probable, or definite) was significantly higher in the heparin group (1.72/1000 line-days) than in both the vancomycin-heparin group (0.37/1000 line-days) and the vancomycin-heparin-ciprofloxacin group (0.55/1000 line-days, p=0.004 and 0.005, respectively). The time to infection was also increased by using an antibiotic flush solution compared with heparin alone (p=0.011 and 0.036 for vancomycin-heparin and vancomycin-heparin-ciprofloxacin, respectively). In addition, of 11 gram-negative infections, 10 occurred in the heparin flush group and 1 occurred in the vancomycin-heparin-ciprofloxacin group.

There were no cases of vancomycin-resistant organisms recovered from sterile sites during the study period. Peripheral ciprofloxacin blood levels drawn randomly in seven patients after vancomycin-heparin-ciprofloxacin flush were lower than the assay sensitivity limit (< 0.05 µg/ml). Nonneutropenic patients developed significantly fewer catheter-related infections with preventive antibiotic flush solutions than with heparin alone; this protective effect was less dramatic in neutropenic patients. The authors concluded that the use of a vancomycin-heparin-ciprofloxacin or a vancomycin-heparin flush solution significantly reduced the complications associated with tunneled CVCs in immuno-compromised children and could save significant health care resources by reducing the number of catheter-related infections by approximately 70%. Although this was a well-designed study, it is limited by its applicability to patients with totally implantable port devices.

Another group evaluated the antibiotic stability of lock solutions used for prolonged dwell times in 10 pediatric oncology patients with implantable ports.[55] Children were excluded if they required an infusion sooner than 48 hours after antibiotic lock therapy, had a port infection or acute illness, or had current or anticipated antibiotic therapy. Vancomycin or ceftazidime was combined with heparin and locked in each port for 2-34 days, depending on the need for port access. When the port was accessed, up to 1 ml of fluid was aspirated and assayed for antibiotic concentration (40 samples). All measured vancomycin concentrations were greater than 130 µg/ml, and no correlation was noted between concentration and dwell time for vancomycin. However, an inverse correlation was noted between the ceftazidime dwell time and concentration (p<0.0001), which most likely reflects loss of ceftazidime activity. Overall, the concentrations of both antibiotics remained above the MIC90 of susceptible organisms for prolonged periods of time. The authors concluded that the high antibiotic concentrations used in lock solutions were stable for at least 2 weeks in an implantable port lumen and may allow for treatment of port infections. However, this study did not evaluate the effect of antibiotic lock solutions on preventing or treating port infections and was small in nature.

Treatment Trials. In one observational study, researchers evaluated antibiotic lock solutions to treat catheter-related sepsis in 11 patients receiving home parenteral nutrition ( Table 4 ).[45] A total of 24 cases of catheter-related sepsis occurred in 18 catheters over a cumulated 286-month period. After meeting specified criteria for catheter-related sepsis, patients were treated with antibiotic lock alone (group 1, 11 cases) or with a short course of systemic antibiotics followed by an antibiotic lock (group 2, 11 cases). Treatment was chosen according to physician judgment and was not randomly selected. Lock solutions used depended on the infecting organism. The first two cases of catheter-related sepsis were treated with only systemic antibiotics for 3 weeks and served as a basis for comparison.

Of the 18 catheters inserted, 7 catheters were removed owing to migration or blockage (four catheters), septicemia (one catheter), or failed therapy (two catheters, one patient in each group both due to Candida infections after primary bacterial eradication). Successful treatment occurred in 90% of cases of catheter-related sepsis in groups 1 and 2 without catheter removal. Hospital stay was significantly shorter for patients in group 1 than for those in group 2 (average of 4.4 and 7.2 days, respectively, p<0.02) and was approximately 3 times shorter than for patients treated with systemic therapy alone. In addition, negative in-line blood cultures were obtained after 3.8 days in group 1 compared with 4.2 days in group 2. The researchers concluded that antibiotic lock therapy alone can success-fully treat catheter-related sepsis. This study is severely limited by its small sample size, lack of treatment randomization and standardization, and discontinuation of parenteral feeding while the patient was febrile.

In 1990, the same researchers published a second observational study involving antibiotic lock use in 19 patients receiving parenteral nutrition, with 27 episodes of catheter-related sepsis.[57] The study methods were similar to those of their previous study.[45] In this study, however, antibacterial activity of the solutions against E. coli and S. epidermidis was determined with a daily bioassay after storage in syringes at 4°C. Amikacin maintained antibacterial activity for 17 days, whereas vancomycin and minocycline retained activity for 3 and 4 days, respectively. In 20 cases of catheter-related sepsis, antibiotic lock was used alone compared with 7 episodes of catheter-related sepsis treated initially with systemic antibiotics followed by an antibiotic lock. As a result of previously documented incompatibilities between antibiotics (specifically aminoglycosides) and heparin,[67,68] heparin was given as a daily injection in all patients before each nutrition infusion. Gram-positive bacteria were the infecting pathogens in 22 cases, whereas gram-negative organisms were isolated in 5 cases. Two catheters were removed during the study period: one due to S. epidermidis after lock therapy alone and the other due to Enterobacter agglomerans after systemic and antibiotic lock treatment.

The average infection-free time for the remaining 25 episodes was 152 days (range 7-570 days). Of these episodes, 15 were recurrent infections of the same catheter that became reinfected an average of 2 times (range 1-3 times). However, 7 of the 15 recurrent episodes were caused by a different bacterial strain than previously isolated. The average time between recurrent episodes was 2 months.

Combining these results with those from the researchers' previous trial,[45] no significant difference was noted in cure rates between using systemic antibiotics before lock therapy (16/18, 89%) and antibiotic lock technique alone (29/31, 94%). The researchers concluded that antibiotic lock technique was an effective treatment for bacterial catheter-related sepsis not associated with infection of the entry point or the subcutaneous tunnel site. However, it must be kept in mind that these were not randomized trials and physician bias in selecting lock therapy alone versus antibiotic lock plus systemic therapy may have influenced success rates.

Another group studied the use of antibiotic locks for catheter-related sepsis in 36 adult patients with renal failure requiring hemodialysis.[54] Patients were prospectively studied for 12 months after placement of double-lumen central venous catheters. Patients who appeared to be septic were treated empirically with intravenous vancomycin 1 g/week. If blood cultures grew gram-negative organisms, therapy was changed to intravenous ciprofloxacin 200 mg admin-istered at the end of each hemodialysis session.

Half of each antibiotic dose was infused through each catheter lumen for a 15-day treatment period. Patients were excluded if they required catheter removal for phlebitis and/or insertion-site infection, deteriorated clinically, were suspected to have pulmonary embolism or bacterial endocarditis, had persistent bacteremia after 48-72 hours of antibiotic therapy, or were fungemic. In between hemodialysis sessions, either vancomycin or ciprofloxacin 100 µg/ml in 5% sodium heparin was injected into each catheter lumen and allowed to dwell until the following hemodialysis session.

Only 30% of patients developed catheter-related sepsis (13 episodes in 11 patients), and none required catheter removal as a result of persistent infection. The causative organisms were S. epidermidis and P. aeruginosa in 77% of the cases, whereas S. aureus was isolated in 15.4%. The researchers speculated that success could be attributed to the "high concentration of the antibiotic in the endoluminal perfusion," or antibiotic lock technique. However, this was a small, observational study with no control group to fully assess the effectiveness of antibiotic locks compared with systemic therapy alone.

In another study, 11 pediatric patients with 12 CVC infections were treated with antibiotic lock therapy.[62] Patients received antibiotic locks if bacteria were isolated from blood cultures drawn through a catheter after 96 hours of systemic treatment, if blood cultures were sterile from a peripheral venous site, or if they had no acute illness and no focus of infection other than the catheter. The choice of antimicrobial used for the lock solution was based on the isolated organism and was used for 10-14 days. Systemic administration of antibiotics and the use of heparin were prohibited during the antibiotic lock treatment period. Blood cultures were obtained on a daily basis from the catheter to determine the efficacy of the antibiotic lock.

Of the 12 episodes, 10 cases of catheter-related sepsis were cured with antibiotic lock technique. The two treatment failures occurred because of catheter occlusion by a blood clot and/or slime produced by S. epidermidis. The authors stated that antibiotic lock therapy is possibly a better choice for first-line therapy than are systemic antibiotics for treating a catheter infection that has not progressed to systemic infection. However, this was another small, observational, uncontrolled trial in which it is unclear if CVCs were simply colonized or infected.

An open, uncontrolled study of antibiotic lock technique and systemic antibiotics for catheter-related sepsis was performed in patients receiving home parenteral nutrition.[47] Patients were excluded if they were 18 years or younger or if any of the following were present: sepsis, infection with S. aureus, metastatic infection, or catheter tunnel infection. After exclusion due to sepsis or S. aureus infection (three patients), there were nine episodes of catheter-related bacteremia or fungemia treated with intraluminal antimicrobials ( Table 4 ). Antibiotic lock therapy was used for 1-2 weeks with or without up to 1 week of systemic antibiotic therapy. Fungal infections were treated with intraluminal amphotericin B for 15 or more days.

All seven episodes of bacterial catheter-related sepsis were cured by using antibiotic lock solutions for an average of 8.6 days without catheter removal, but four episodes were also treated with an average of 2.1 days of systemic antibiotics. Two Candida infections initially cleared with use of lock therapy and systemic antifungals, but both episodes relapsed 6 weeks to 7 months later, one of which required catheter removal. In six patients being treated with antibiotic locks, vancomycin and gentamicin levels were assessed after dwelling for 8-12 hours in the catheters. Both drug concentrations remained above the MICs of commonly infecting organisms (3.822 ± 1.017 µg/ml for vancomycin [five patients] and 3.33 µg/ml for gentamicin [one patient]). The authors concluded that antibiotic locks used for 7-8 days were as effective as standard systemic antibiotic treatment periods of 10-16 days for clearing catheter-related bacteremia. They also stated that vancomycin and gentamicin lock concentrations remained elevated for at least 8-12 hours after being instilled in the catheter. In addition, daily use of amphotericin B lock technique was effective in suppressing Candida infection and extending the catheter life for months.

This study is limited by the small number of episodes of catheter-related sepsis and the lack of a comparison group to assess antibiotic lock technique monotherapy versus systemic therapy. The authors did not elaborate on the decrease in antibiotic lock concentrations found in a small subanalysis in this study.

In another small study, 16 patients with venous access ports who had human immunodeficiency virus or cancer were treated with antibiotic locks for febrile CVC-related bacteremia.[63] Patients were excluded if they had localized insertion- or tunnel-site infection. The antibiotic lock was used in combination with systemic antibiotics ( Table 4 ) directed against the isolated organisms.

Treatment without port removal was successful in five patients (31%), whereas partial response (defined as cure of the initial infection followed by a recurrent infection with a different microorganism) occurred in two patients. Nine patients (56%) failed therapy and required port removal. Although this was a small observational study, it involved patients with venous access ports that contain a reservoir connected to the catheter. The presence of this reservoir can affect the success of an antibiotic lock secondary to the size of the reservoir and presence of "residue," such as fibrin or other proteins, in the lumen. The residue can serve as a harbor for bacteria, limiting the efficacy of antibiotics, which may explain the high rate of failure found in this study.

Another group of investigators examined the use of antibiotic locks to treat catheter-related sepsis in 42 patients receiving home parenteral nutrition through either single-lumen cuffed silicone CVCs (64%) or implantable ports (26%).[64] Based on specific diagnostic criteria, 39 episodes of catheter-related sepsis occurred during the study. The most frequently isolated organism was S. epidermidis, and patients were treated with the appropriate antibiotics for the organism(s) isolated (antibiotics not specified in study). The lock technique was more effective in treating infections of CVCs than implanted ports (p<0.05), reinforcing the results of the above-mentioned study.[63] The investigators concluded that antibiotic lock therapy can be an effective method for salvaging and sterilizing CVCs, but this study is limited by its small, uncontrolled, observational nature and lack of detail regarding the methodology of its antibiotic lock technique.