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


Subcutaneous tunneling reduced catheter-related sepsis in critically ill patients

ACP J Club. 1997 Mar-Apr;126:42. doi:10.7326/ACPJC-1997-126-2-042

Source Citation

Timsit JF, Sebille V, Farkas JC, et al. Effect of subcutaneous tunneling on internal jugular catheter-related sepsis in critically ill patients. A prospective randomized multicenter study. JAMA. 1996 Nov 6;276:1416-20.



To determine the efficacy and safety of subcutaneous catheter tunneling (SCT) on internal jugular catheter-related sepsis.


Randomized controlled trial with follow-up to hospital discharge.


3 intensive care units (ICUs) in Paris, France.


241 patients (mean age 65 y, 72% men) who were admitted to the ICU and were expected to need ≥ 48 hours of catheterization. Exclusion criteria were age ≤ 18 years, catheterization by guide-wire exchange, need for triple-lumen catheter, previous tracheostomy, or contraindicated SCT because of previous neck or intraclavicular surgery. Follow-up was 96%.


Patients were stratified by center and by the number of lumens. Final analysis included 117 patients allocated to SCT and 114 allocated to non-tunneled catheters. SCT used single- or double-lumen tunneled catheters. The nontunneled catheters had the same internal and external diameters. All catheters were inserted under strict aseptic conditions.

Main outcome measures

Catheter-related septicemia, bacteremic catheter-related sepsis, and catheter colonization (≥ 103 colony-forming units/mL).

Main results

Analysis was by intention to treat. Systemic sepsis developed in fewer patients who received tunneled catheters than in patients who received nontunneled catheters ({6% vs 16%}*, P = 0.02). {This absolute risk reduction (ARR) of 10% means that 10 patients would need to be treated (NNT) with SCT (rather than nontunneled catheters) to prevent 1 additional patient from developing systemic sepsis, 95% CI 5 to 53; the relative risk reduction (RRR) was 62%, CI 15% to 83%.}* Similarly, bacteremic sepsis developed in 4 patients (3%) who received tunneled catheters compared with 13 patients (11%) who received nontunneled catheters (P = 0.02) {ARR 8%; NNT 13, CI 6 to 73; RRR 70%, CI 16% to 89%}*. The groups did not differ for the incidence of catheter colonization (17% vs 25%, P = 0.1 {ARR 8%, CI -2% to 19%}*).


Subcutaneous catheter tunneling, as compared with nontunneling, reduced internal jugular catheter-related sepsis in critically ill patients.

Sources of funding: Plastimed Company, Saint Leu, France, and Lederle Company.

For article reprint: Dr. J.F. Timsit, Clinique de Réanimation des Polyvalente, Hôpital St. Joseph, 185, rue Raymond Losserand, 75674 Paris, France. FAX 331-44-123-280.

*Numbers calculated from data in article.


Central venous catheter infections can come from several sources; tunneling is used to reduce those that originate from the skin. This study focused on the internal jugular site, whereas previous trials have evaluated tunneling almost exclusively in patients with subclavian catheters. Although no randomized trial comparing subclavian with internal jugular insertion has been published, cohort studies suggest that infection rates are higher in patients with internal jugular catheters. The 12 randomized trials that compared insertion of subcutanian venous catheters using SCT with insertion using standard placement had inconsistent results.

The design of the trial by Timsit and colleagues was strong: Randomization was concealed, groups had similar illness severity and demographic characteristics, and follow-up was nearly complete. Interventions that may also affect rates of infection (dressing and use of antimicrobial agents) were similar between groups. Methods to minimize diagnostic suspicion bias in this unblinded study included 1) culturing and catheter removal done according to a prespecified, explicit protocol that was used by caregivers other than the investigators and 2) independent examination of outcome data by 2 authors who were blinded to treatment group.

In this study, tunneling was associated with lower rates of catheter-related sepsis and septicemia and a trend toward decreased colonization. The first puncture success rate was 75%, but the final success rate was 96%. The arterial puncture rate was 7%, and one third of lines were inserted by a senior physician. Most catheters were double lumen, were in situ for a mean of 8 days, and were used at the discretion of the ICU team. The adoption of this management approach in other settings depends on several issues, such as the reproducibility of the tunneling technique described by the authors and the technical competence of the clinicians who use this method.

Systematic reviews of the myriad management strategies for central venous catheterization are needed. More rigorous randomized trials such as this one will help to clarify areas of controversy and further develop the evidence base for sound practice.

Deborah J. Cook, MD
McMaster UniversityHamilton, Ontario, Canada
Adrienne G. Randolph, MD, MSc
University of TorontoToronto, Ontario, Canada