Antibiotic-Impregnated Catheters Superior

Abstract & Commentary

Synopsis: Central venous catheters (CVCs) impregnated with minocycline and rifampin were one-third as likely to become colonized and only one-twelfth as likely to be the source of a bloodstream infection as CVCs impregnated with chlorhexidine and silver sulfadiazine.

Source: Darouiche RO, et al. N Engl J Med 1999; 340:1-8.

Prevention of nosocomial infection related to central venous access is an important clinical goal. Darouiche and colleagues prospectively compared two antimicrobial-impregnated central venous catheters (CVCs) in a group of 865 insertions at 12 university-affiliated teaching hospitals. Patients who were at high risk (ICU patients or those known to be immunocompromised) and expected to need central access for at least three days were randomly chosen to receive 7 F, triple-lumen CVCs impregnated with chlorhexidine silver sulfadiazine (CSS) or minocycline and rifampin (MR). Only catheters inserted with an initial venous stick were included in the study population; no patients using over-a-wire change were studied. In addition to the usual semiquantitative culture techniques applied to catheter segments on removal, sonication culture of the catheter was used to increase the yield of potentially colonizing organisms. Indications for catheter removal were when it was no longer needed, if mechanical problems occurred, or if catheter-related infection was suspected.

Randomization for patient demographics, duration of cannulation, disease process, and reason for removal was successful in the 738 catheter placements in 698 patients completing the study (85% of the total 865 CVCs initially entered). Only one of the 350 MR CVCs was associated with a bloodstream infection compared to 13 of the 370 CSS catheters (P < 0.002). Twenty-eight of the MR CVCs (7.9%) were colonized as compared to 87 of the CSS (22.8%); P = 0.001. No complications related to CVC impregnation, including patient hypersensitivity or antibiotic resistance, were noted in either group.


Both of the catheters used in this study have previously been shown to reduce the incidence of catheter-related infections and to about the same extent when compared to nontreated catheters. This is the first head-to-head trial of both types of catheter. Darouiche et al mention several facts about the devices they used: the MR catheters are impregnated inside and outside; the catheters used in this study had almost five times the amount of minocycline and 20 times the amount of rifampin per catheter as those used in earlier studies; and CSS catheters are only impregnated on the outer surface. The MR catheters are about $11 more than the CSS and about $50 more than a conventional catheter (using my estimate of $20 for a conventional CVC).

Other interesting factors reported in this study were that the femoral and jugular routes (together) had two- to fivefold increases in catheter-related infection rates compared to subclavian placement; that men were more likely to experience catheter-related infection than women; that the rate of catheter-related infection was related to the duration of catheterization; and that patients in the ICU or on mechanical ventilation had higher catheter-related infection rates.

While the outcome differences are impressive, there are several concerns about these study data and their implications. The infection rate of the CSS catheters reported in this study is much higher than previously reported. The average reported bloodstream infection rate is only 1%, with a colonization rate of around 12% for CSS catheters in other studies. The rate of bloodstream infections of 3.4% and a colonization rate of 22% suggest a problem with diagnostic accuracy or catheter care. Darouiche et al suggest that sonication enhances the "yield" of organisms and may be part of the reason for the high rates. This procedure is not yet the accepted method of diagnosis and is not routinely available for clinical use. The data should have included what was found using the traditional rolled-catheter, colony-forming units method of identifying infection.

Another explanation of the results offered by Darouiche et al is that the MR catheters are coated inside as well as outside. If line infection originates hematogenously as well as from the skin, this may be important in decreasing infection. It also suggests that failure of aseptic line management may be occurring in this study. Line dressing techniques also contribute to catheter-related infections. The dressing technique is not described in this study. Occlusive dressings and use of neosporin ointment have the highest catheter-related infection rates; nonocclusive, chlorohexidine cream, the lowest. The type of line dressing should have been made clear in the paper.

These concerns aside, this is promising new information and any patient with a central venous catheter that is expected to remain in place for more than seven days may benefit from an antimicrobial catheter or placement technique. The true difference between MR and CSS catheters, catheter tunneling, and collagen plugs awaits additional testing.