Fragou et al randomized 401 ICU patients to undergo subclavian vein central line placement with guidance either by ultrasound or anatomic landmarks. All cannulations were with an infraclavicular approach. The ultrasound group had a higher success rate (100% vs 87.5%), shorter time to access and fewer number of attempts, and a lower rate of complications. (n=401) Crit Care Med 2011;39(7):1607-1612.

Kudos to the authors for reporting accurate and scary complication rates using the time-honored landmark method: 5% pneumothorax, 4% hemothorax, 3% brachial plexus injury, and one case of cardiac tamponade. My spider-sense tells me, these were evenly distributed enough that each individual doc could continue to claim "I've done HUNDREDS of these, and had ONE complication!" Paging Peter Pronovost ...