Web site designed and maintained by Chris Cairns  © SICS EBM Group 2004                                  

Subclavian vs. Internal Jugular CVC: A meta-analysis

Citation/s: Ruesch S, et al. Complications of central venous catheters: Internal jugular versus subclavian access - A systematic review. Crit Care Med 2002; 30:454-460.

Lead author's name and fax: Sibylle Ruesch, Department of Anaesthesiology, University Hospital, Geneva, Switzerland.

Three-part Clinical Question: In all patients with a non-tunnelled CVC line, which site: subclavian or internal jugular is associated with fewer complications? For the purposes of this CAT we will concentrate on this papers investigation of infective complications.
 

Search Terms: see summary

The Review:

Data Sources:  Cochrane Library, Medline, Embase
Study Selection: Up until June 30th, 2000. Key words: central venous catheter, catheterization, catheterisation, subclavian, jugular, complication, infection, thrombosis, success rate, stenosis, pneumothorax, haemothorax, clinical trial, prospective. Included if: full reports of int Jug. Vs Fem, adults or children, dichotomus data, peer reviewed, any type of CVC. Excluded if letters, abstracts, reviews, animal studies, post mortem studies, tunnelled lines, implantable devices. Studies were also excluded if there was a greater that twofold difference in the number of subclavian to int. jugular catheters (the authors did this in an attempt to reduce selection bias).
Data Extraction: Patients, clinical settings, operators.
The studies were multiple independent reviews of individual reports. They were tested for heterogeneity.

The Evidence:

EBM questions:

 1.  Do the methods allow accurate testing of the hypothesis? No. It was unusual that studies with a greater than two fold ratio of different sites were excluded (at least 7 studies). One would have thought that a properly conducted meta-analysis (with tests for heterogeneity) would have made this exclusion criteria unnecessary. Only abstracts were screened in the initial stages which may partly explain the next point. The search strategy was week: two large observational studies were missed by the literature search: Goetz et al, Infect Control Hosp Epidemiol 1998;19:842-845, Richet et al, J Clin Micro 1990;28:2520-2525 (over 1000 CVCs in these two studies – only 707 catheters in this meta-analysis) as well as several RCTs of other interventions which include information on different site infection rates. All these points suggest that methodology was poor

2.      Do the statistical tests correctly test the results to allow differentiation of statistically significant results? Yes but heterogeneous population.

3.      Are conclusions valid in light of the results? No. The authors claimed the difference in CVCrBSI, although not statistically significant, was clinically significant. This is an inaccurate conclusion. The study is sufficiently flawed to disregard the conclusions.

4.      Did results get omitted, and why? No

5.      Did they suggest areas of further research? Yes. RCTs of different sites.

6.      Did they make any recommendations based on the results and were they appropriate? No.

7.      Is the study relevant to my clinical practice? No. Should be ignored.

8.      What level of evidence does this study represent? 1^-

9.      What grade of recommendation can I make on this result alone? none

10.  What grade of recommendation can I make when this study is considered along with other available evidence? None.

11.  Should I change my practice because of these results? No.

12.  Should I audit my current practice because of these results? No.

Appraised by: Dr Chris Cairns, Consultant, ICU, Stirling Royal Infirmary, UK ; 01 June 2005

Email: Chris.Cairns@fvah.scot.nhs.uk

Kill or Update By: June 2010

Printer friendly version

©SICS EBMG 2005