Locking out Colonization of Central Venous Catheters with Antibiotics

Abstract & Commentary

Synopsis: The use of antibiotic locks to sterilize colonized central venous catheters may be justified, although the evidence in favor of the technique is anecdotal and otherwise flawed.

Source: Berrington A, Gould FK. Use of antibiotic locks to treat colonized central venous catheters. J Antimicrob Chemother. 2001;48:597-603.

Berrington and Gould review the limited data on the use of antibiotic locks as a means of treating colonized central venous catheters (CVCs) and conclude that there might, indeed, be circumstances in which the use is justified even though the level of evidence is low. The idea is to salvage catheters in patients such as hematopoietic stem cell transplant recipients for whom it is vital to maintain venous access by instilling antibiotics into the catheter lumen just enough to fill in (usually 1-2 mL) and then closing the device off for a number of hours. The aminoglycosides, gentamicin and amikacin, have been used at 5-13 mg/mL and 2-40 mg/mL, respectively, and vancomycin has also been used mainly in settings where colonization by coagulase-negative staphylococci is commonplace at concentrations of 1-80 mg/mL. This approach has been reported in at least 13 studies, though it was used in conjunction with systemic treatment in 6 of them and heparin was included in the lock fluid in 3 studies. There was no uniformity in adminstering the lock, nor in terms of how long the treatment was given, varying from 1 to 2 weeks. Nonetheless, Berrington and Gould cautiously concluded that although the ideal is always to remove a colonized catheter, this may run into conflict with other more pressing clinical imperatives and, so, the balance of risks might favor short-term antibiotic lock treatment either alone for patients without concurrent bacteremia or as an adjunct to systemic therapy when bacteremia is present.

Comment by J. Peter Donnelly, PhD

The antibiotic lock technique does seem a rather good idea for preserving the life of CVCs since, if effective, it would eliminate the need for seeking an alternative means of accessing the bloodstream for a substantial number of patients and would prove much cheaper and probably safer. Moreover, leaving the antibiotic in contact with the bacteria stuck to the inner surface of the catheter is more likely to exert its effect than the ephemeral contact resulting from standard infusions of the same drug. However, there is really no evidence that it works any better than, say, flushing the device out with heparin or even saline, as there are no randomized controlled trials. It also seems unlikely that the concentration of antibiotic sitting in the lumen would remain at the original level for long since drug would leach into the circulation driven by the law of mass action. Again, there are no data. Neither are there any data to indicate which concentration of drug should be used, how long it should be left in contact, and how long should the treatment go on. There are also issues of physical interactions that need to be addressed. Might not disinfectants such as ethanol be just as effective and might some of the nonantimicrobial drugs interfere with the action of the antibiotic? We simply do not know.

These days multilumen devices are commonplace, making the mechanics of the technique more complex. A patient being managed with a triple-lumen catheter from which blood taken through each lumen yields the same bacterium would need a lock on each lumen to clear the organism. Is each lumen to be left undisturbed for several hours each day to allow the treatment to work? This seems implausible given the patients involved. Yet, the fact that a review article on this subject has appeared in a leading journal suggests that the use of antibiotic locks continues to accelerate for pragmatic rather than scientific reasons or simply because of the recent advice of the IDSA, which declared that "antibiotic lock therapy should be used for 2 weeks together with standard systemic therapy for treating catheter-related bacteremia due to S aureus, coagulase-negative staphylococci, and gram-negative bacilli for suspected intraluminal infection of tunnelled devices in the absence of tunnel or pocket infection." In short, there are many more questions than answers about this deceptively simple and intuitively attractive approach, but proving the effectiveness of the technique will be no easy matter given the huge number of variables involved and the confounding factors of other parenteral antimicrobial agents that may be given for other reasons. As with many interventions brought to life to meet an unmet need, only time will tell.

Dr. Donnelly, Clinical Microbiologist, University Hospital, Nijmegen, The Netherlands, is Associate Editor of Infectious Disease Alert.