Stay up-to-date in pulmonary and critical care. No spam.
Where to place a central venous catheter is a decision driven mainly by individual experience and preference. The limited evidence available has not established any site as superior; the subclavian position has been reported as being less infection-prone, but more likely to cause pneumothorax, compared to other sites. A large French randomized trial adds significantly to the evidence base.
Authors of the 3SITES study randomized more than 3,000 patients in France requiring central venous catheterization to have their line placed in either the internal jugular, subclavian, or femoral position.
To increase power (i.e., to reduce the number of enrolled patients needed to confidently find an effect), the primary outcome was a composite of either a catheter related bloodstream infection, or a deep venous thrombosis at the site.
Ultrasound was only used two-thirds of the time in internal jugular line placements (vs. 16% in subclavian line placements).
Operators were experienced, with 50+ central line placements under their belts (although the lines could also be placed by supervised inexperienced operators).
The subclavian position was by far the superior site for infections or DVTs, with less than half the composite rate (1.5 events per 1,000 catheter days) compared to the internal jugular (3.6) or femoral (4.6).
However, the patients undergoing subclavian vein puncture had three times the rate of pneumothorax requiring chest tubes for drainage (1.5% vs. 0.5%) compared to the patients with internal jugular vein cannulations.
That's a number needed to harm of 100 patients with subclavian CVL placements, to cause an additional pneumothorax. It's also a 1 in 67 chance of pneumothorax (subclavian) vs. 1 in 200 (internal jugular). Rates of pneumothorax by less-experienced operators outside this study would be expected to be higher.
Despite the widespread taboo against the femoral position for central line placement, rates of CLABSI were no higher for femoral central venous catheters (1.2%) than for the internal jugular position (1.4%).
However, femoral catheters produced symptomatic DVT 1.4% of the time, vs. 0.9% for the IJ position and 0.5% subclavians.
Summing Up: What's the "Best" Site for Central Line Placement?
Using the subclavian site for central venous catheterization reduced infections and DVT to a minimum, but tripled the risk of pneumothorax compared to the internal jugular (IJ) position. Using the femoral position eliminated pneumothorax risk, and was comparable to the IJ in infection risk, but significantly increased DVT risk.
These data suggest that there is probably no ideal position for central venous catheterization. Pneumothorax is more feared, dramatic, and blame-laden than CLABSI or DVT, leading most intensivists and midlevels to prefer internal jugular placement. Deaths attributable directly to central line placement (by either pneumothorax or CLABSI) are too infrequent to consider one catheter position definitively safer.
All supervisors in this trial had placed >50 central lines -- but the operator (person actually placing the line) could be an inexperienced trainee. Complication rates should be expected to be higher by less-experienced operators, especially if unsupervised.
It's impossible to know whether universal use of ultrasound for the internal jugular line placements, and chlorhexidine antiseptic skin cleansing for all patients, would have changed the observed complication rates.
All central venous catheters should be considered for removal every day, and removed as soon as feasible. Most patients who are not receiving vasopressor infusions do not require a central venous catheter.
Intravascular Complications of Central Venous Catheterization by Insertion Site. N Engl J Med. 2015 Sep 24;373(13):1220-9.
What is the best site for central venous catheter insertion in critically ill patients? Crit Care. 2003; 7(6): 397–399.
The incidence of infectious complications of central venous catheters at the subclavian, internal jugular, and femoral sites in an intensive care unit population. Crit Care Med. 2005 Jan;33(1):13-20; discussion 234-5.