Current issues of ACP Journal Club are published in Annals of Internal Medicine


Scheduled replacement of central vascular catheters

ACP J Club. 1994 Mar-April;120:55. doi:10.7326/ACPJC-1994-120-2-055

To the Editor

In the March/April 1993 issue, Kruse (1) favorably reviewed an article by Cobb and colleagues (2) that challenged the widely held belief that increased infection rates result from leaving central vascular catheters (CVCs) unchanged for longer than 3 to 4 days. In a randomized, controlled trial of scheduled replacement of central venous and pulmonary artery catheters, Cobb found more bloodstream infections and mechanical complications "pneumothorax, arrhythmia, suspected thrombosis, and bleeding" in patients on a fixed, every 3-day replacement schedule compared with patients whose lines were changed only for such clinical indications as unexplained fever or CVC malfunction. Kruse noted with approval the article's recommendation that the conventional, scheduled approach to line changing be abandoned in favor of leaving CVCs in place until the development of clinical indications for replacement.

We believe that certain limitations in the study's design make acceptance of its recommendations premature. The most important limitation is that all of the catheterizations were performed by first- and second-year residents. In community hospitals, the job is done by surgeons, cardiac or pulmonary specialists, or experienced internists, whose rates of complications are likely to be considerably lower than those of physicians in their earliest years of training (3, 4). Thus, the rates of infectious and mechanical complications observed in the study by Cobb and colleagues do not provide a secure basis for decision making outside the teaching hospital environment.

A second important weakness in the study is the exclusion of 17% of the patients after randomization. This is problematic because the causes for exclusion were unevenly distributed between the study groups. In addition, the study failed to record the outcome of excluded cases. Nonrandomly excluded cases could bias the results in either direction, depending on the composition of the excluded group and its rate of complications. The trial's results should have been analyzed by “intention to treat,” in which, after randomization, groups are compared using all the patients initially assigned to each group even if they did not receive the treatment they were assigned to undergo. This approach would have minimized the chance of a falsely optimistic appraisal of the benefit of leaving CVCs in place until there is a clinical indication for replacement.

Amnon Lahad, MD
University of Washington
Seattle, Washington

Jaime Friedman, MD
Sand Point Internists
Seattle, Washington

In response

Drs. Lahad and Friedman raise concern about the generalizability of our study because of high complication rates (5). Although complication rates may be somewhat higher in teaching hospitals in general, the rates of specific CVC complications in our study (e.g., bacteremia in 1.5% and pneumothorax in 1.1% of patients with CVCs) were lower than those reported in most other studies. In any event, the key issue is not the absolute incidence of complications but the relative risk for complications with scheduled replacement as compared with change as clinically indicated, which should apply in both teaching and nonteaching hospitals.

The other concern raised by Lahad and Friedman relates to exclusion of “17% of patients after randomization.” Almost half of this 17% (i.e., 8%) did not meet randomization criteria because they did not require a CVC for more than 72 hours. For the other exclusions, outcome data on catheter infection were unavailable because the catheter was lost in transit to the laboratory or grossly contaminated and discarded. None of these patients had a positive blood culture, excluding the possibility of CVC-related bloodstream infection, but the CVC colonization rate could not be calculated for lost catheters. The study groups remained comparable after exclusion of these patients.

Randomized trials have tested the hypothesis that scheduled CVC replacement every 3 days (5, 6) or 7 days (7) would result in a lower rate of infection. Because the incidence of catheter infection does not change between days 3 and 7 of catheterization (5, 8, 9) and because of the similar designs and homogeneity of results of these studies, we have pooled the data from these 3 trials with the following results: 27% of 211 patients having scheduled replacement had colonized catheters compared with 14% of 187 patients having unscheduled replacements. 11% with scheduled replacement had a catheter-related bloodstream infection as compared with 10% of patients having unscheduled changes.

Inserting a new CVC every few days is, at best, a useless and expensive practice, which provides a false sense of safety. It increases risk for mechanical complications, such as pneumothorax and hemothorax with new punctures, and for infections with guidewire exchanges (5, 10-12). For the minority of patients in intensive care who require a CVC for more than a few days, a silver-impregnated collagen cuff (13, 14) or an antiseptic impregnated catheter (15) can be used to decrease the risk for bloodstream infection.

Barry M. Farr, MD, MSc
University of Virginia
Charlottesville, Virginia


1. Kruse JA.Scheduled replacement of central vascular catheters. ACP J Club. 1993 Mar/Apr:47 (Ann Intern Med. vol 118, suppl 2).

2. Cobb DK, High KP, Sawyer RG, et al. A controlled trial of scheduled replacement of central venous and pulmonary-artery catheters. N Engl J Med. 1992;327:1062-68.

3. Sitzmann JV, Townsend TR, Siler MC, Barllett JG. Septic and technical complications of central venous catheterization. A prospective study of 200 consecutive patients. Ann Surg. 1985;202:766-70.

4. Venkataraman ST, Orr RA. Percutaneous infraclavicular subclavian vein catheterization in critically ill infants and children. J Pediatr. 1988;113:480-5.

1. Cobb DK, High KP, Sawyer RG, et al. A controlled trial of scheduled replacement of central venous and pulmonary-artery catheters.N Engl J Med. 1992;327:1062-8.

2. Powell C, Kudsk KA, Mandelbaum JA, Fibri PJ. Effect of frequent guidewire changes on triple-lumen catheter sepsis. J Parenter Enteral Nutr. 1988;12:462-4.

3. Eyer S, Brummitt C, Crossley K, Siegel R, Cerra F. Catheter-related sepsis: prospective, randomized study of three methods of long-term catheter maintenance. Crit Care Med. 1990;18:1073-9.

4. Stenzel JP, Green TP, Fuhrman BP, et al. Percutaneous central venous catheterization in a pediatric intensive care unit: a survival analysis of complications. Crit Care Med. 1989;17:984-8.

5. Widmer AF. In: Wenzel RP, ed. Prevention and Control of Nosocomial Infections. 2d ed. Baltimore: Williams & Wilkins, 1993:562.

6. Newsome HH Jr, Armstrong CW, Mayhall GC, et al. Mechanical complications from insertion of subclavian venous feeding catheters: comparison of de novo percutaneous venipuncture to change of catheter over guidewire. J Parenter Enteral Nutr. 1984;8:560-2.

7. Michel LA, Bradpiece HA, Randour P, Pouthier F. Safety of central venous catheter change over guidewire for suspected catheter-related sepsis. A prospective randomized trial. Int Surg. 1988;73:180-6.

8. Snyder RH, Archer FJ, Endy T, et al. Catheter infection. A comparison of two catheter maintenance techniques. Ann Surg. 1988;208:651-3.

9. Flowers RH 3d, Schwenzer KJ, Kopel RF, et al. Efficacy of an attachable subcutaneous cuff for the prevention of intravascular catheter-related infection. A randomized, controlled trial. JAMA. 1989;261:878-83.

10. Maki DG, Cobb L, Garman JF, et al. An attachable silver-impregnated cuff for prevention of infection with central venous catheters: a prospective randomized multicenter trial. Am J Med. 1988;85:307-14.

11. Maki DG, Wheeler SJ, Stolz SM, Mermel LA. In: Program and Abstracts of the 31st Interscience Conference on Antimicrobial Agents and Chemotherapy. Chicago, 29 September to 2 October, 1991. Washington, D.C.: American Society for Microbiology; 1991:176 [Abstract].