Objective: To compare ultrasound (US)-guided vs. landmark-guided techniques for central venous access (CVA) in the emergency department.
Methods: This was a prospective study of consecutive patients enrolled at a university teaching hospital with an annual census of approximately 100,000. On even days patients had CVA with ultrasonic assistance; patients presenting on odd days had CVA via traditional landmark techniques. Ultrasound users were emergency medicine faculty or residents who completed a one-hour training session. A data collection tool with 17 variables was completed for each central line placed. Variables were compared using the independent t-test, Fisher's exact test, and the non-parametric Mann-Whitney U test.
Results: Between August 1, 2000, and February 1, 2001, data for 122 subjects (n = 51 for US, and n = 71 for landmark) were collected. Variables with statistically significant differences are as follows. Mean (+/-SD) time from skin puncture to blood flash was 115 (+/-184) seconds for the US group vs. 512 (+/-698) seconds for the landmark group (p < 0.0001). The mean number of CVA attempts in the US group was 1.6 (+/-1.0) vs. 3.5 (+/-2.7) in the landmark group (p < 0.0001). Acute complications were comparable between groups. Comparisons for time, number of CVA attempts, and complications were done specifically for a subset of patients considered to be "difficult stick" due to predefined criteria regarding body habitus or vascular disease. Patients considered to be "difficult sticks" required a significantly longer amount of time (p < 0.001) for CVA via the landmark technique than patients considered to be "difficult sticks" who had CVA with ultrasonic guidance. Time to line placement for the landmark group was 462.7 (+/-627) seconds vs. 93.3 (+/-176) seconds in the US group. Comparing the same subset also revealed an increase in number of required CVA attempts for the landmark technique group. The number of acute complications in the "difficult stick" patients did not show statistical significance (p = 1.00). The landmark group had 60% "difficult sticks," while the ultrasound group had 80%, although the difference was not statistically significant (p = 0.08).
Conclusions: Emergency physicians with limited training and experience are able to use ultrasound as an adjunct for central venous access. Ultrasound technology may decrease the number of CVA attempts required to cannulate a central vein and will decrease the amount of time required to cannulate the vein starting from the time when the needle is on the skin, after the ultrasound machine has been set up and turned on. These results are especially true for those patients considered to be "difficult sticks."