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


Scheduled replacement of central venous catheters did not reduce the incidence of infection or complications

ACP J Club. 1993 Mar-April;118:47. doi:10.7326/ACPJC-1993-118-2-047

Related Content in the Archives
Scheduled Replacement of Central Vascular Catheters

Source Citation

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 Oct 8;327:1062-8.



To compare the incidence of infections and mechanical complications in patients in the intensive care unit (ICU) who have central venous and pulmonary-artery catheters changed according to 1 of 4 schedules.


Randomized controlled trial with follow-up to discharge from the ICU.


ICUs of a large U.S. teaching hospital.


Adult patients who required triple-lumen or pulmonary-artery catheters inserted through the internal jugular or subclavian veins for > 72 hours. Exclusion criteria were age < 18 years, coagulopathy or thrombocytopenia, catheters that had not been inserted aseptically, or catheters previously placed by guide-wire-assisted exchange. 523 catheters in 160 patients (mean age 58 y, 57% men, mean APACHE II score 19) were included.


Patients were assigned to 4 catheter replacement groups: group 1, every 3 days with a new puncture site (n = 35); group 2, every 3 days with guide-wire-assisted exchange (n = 40); group 3, as clinically indicated with a new puncture site (n = 41); and group 4, as clinically indicated with guide-wire-assisted exchange (n = 44). Clinical decisions to replace the catheters were made by primary physicians and based on exit-site infections, unexplained fever, and catheter malfunction. Insertion sites were examined and dressings changed every 48 hours. Insertions were done aseptically and checked by radiograph.

Main outcome measures

Incidence of catheter colonization or blood-stream catheter-related infections, mechanical complications (pneumothorax, temporary arrhythmia requiring intervention, bleeding requiring transfusion, and clinical diagnosis of central venous thrombosis).

Main results

The groups did not differ for mortality (range, 11% to 20%), and no patient was believed to have died of a catheter-related complication. The groups did not differ for rates of bloodstream infections per patient or per catheter, and rates of catheter colonization trends, although nonsignificant, favored replacement when clinically indicated. For mechanical complications, 11 of 14 (79%) were associated with a new puncture site, and 3 (21%) were associated with guide-wire-assisted exchange (P = 0.005).


Routine replacement of central venous catheters every 3 days did not prevent infections. When replacement is clinically indicated, the decision to replace catheters over wire (increased risk for infection) or through new puncture sites (increased risk for mechanical complications) has to be made for each patient individually.

Source of funding: Not stated.

For article reprint: Dr. B.M. Farr, Department of Medicine, Box 473, University of Virginia Health Sciences Center, Charlottesville, VA 22908, USA. FAX 804-243-6483.


Routine replacement of central vascular catheters after a certain number of days is recommended and practiced by many physicians caring for patients who are critically ill. The conventional wisdom is that this approach may decrease the rate of catheter-related infections, albeit at the expense of increasing the rate of insertion-related complications. The study by Cobb and colleagues questions the former rationale and confirms the latter.

Like most clinical trials of catheterization complications, this one is limited by sample size considerations. Although the total number of patients and catheterizations seems large, there were only 8 catheter-related bloodstream infections. The study was designed to have sufficient statistical power to detect a difference in infection rates per catheter comparing scheduled catheter replacement and clinically indicated exchanges only if the difference between the groups was at least 10% (e.g., 5% vs. 15%). In other words, differences less than this could exist and not be detected by this investigation. The study found identical infection rates of 1.5% in each group. Whether there is actually a difference in infection rates between these 2 methods of catheter replacement is indeterminate from the sample size used in this trial. Subgroup analysis is even more constrained by the low rate of complications and overall sample size.

Nevertheless, the authors point out that there is no hard evidence to support routinely changing central vascular catheters on a scheduled basis. Further, their results at least suggest that any difference in the rate of infectious complications using routine changes compared with clinically guided replacement is likely to be small. On the other hand, they have clearly shown that the rate of mechanical complications is greater with scheduled replacement. Taken together, this information cogently argues for a re-evaluation of the practice of routine line changes.

James A. Kruse, MD
Wayne State University School of MedicineDetroit, Michigan, USA