• 
• 

Three Part Question

In [patients undergoing elective cardiac surgery, requiring a central line] does the use of the [ultrasound guided central line placement technique] reduce [time to insertion and incidence of complications]

Clinical Scenario

You are a consultant anaesthetist inserting a central line in a patient undergoing elective coronary artery bypass grafting (CABG). A junior colleague asks you why you are not following the NICE guidelines which state that you should always use an ultrasound probe when placing a central line. You reply that you do more than 250 central line placements per year and that your complication rate is very low and that an ultrasound scan would just slow down the start of the operation. Your junior colleague then asks "if you did have a complication, couldn't a lawyer criticise you for not using an ultrasound probe in all your cases?" You are unsure how to respond to this and therefore you resolve to look up the evidence for use of ultrasound guided central line placement in elective cardiac surgical patients.

Search Strategy

Medline 1966-March 2004 using the OVID interface
[exp ultrasonography/ OR ultrasound.mp] AND [exp Catheterization, central Venous/ OR central venous catheter.mp OR central line.mp] AND [cardiac surgery.mp OR cardiac operation$.mp OR exp cardiopulmonary bypass/ OR cardiopulmonary bypass.mp]

Search Outcome

193 papers were found from Medline using the reported search of which 5 papers presented the best evidence to answer the clinical question. This included one meta-analysis, which was used to formulate the NICE guidelines, which we also reviewed. These papers are presented in Table 1.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Hind et al, 2003, UK	18 trials (1646 patients) studying either Doppler or 2-D ultrasound guided central line placement compared to landmark method in a range of clinical settings Searched 15 bibliographic databases until Oct 2001, as part of the NICE guideline development process	Systematic review and meta-analysis (level 1a)	Incidence of complications in RIJ vein cannulation in adults	Relative risk of complication using 2D ultrasound is 43% less than the risk using landmark method (CI 22% to 87%)	Very wide range of patient groups included including cardiac arrest, chemotherapy, dialysis and infant groups. This is a BMJ report of the NICE guidelines technical report No49, although 2 papers were removed from those reviewed by NICE as they were in abstract form only.
			Failed catheter placement (Adults having RIJ placement)	Ultrasound group: 5/296 (1.7%) Landmark group 68/312 (22%) Odds ratio 0.14 (CI 0.06 to 0.33)
			Time to catheter placement	Average 69 seconds quicker to place central line with 2-D ultrasound probe (CI 46 to 92 seconds) Average 34seconds slower to place central line with Doppler probe (CI from 54 seconds quicker to 124 secs slower)
Sulek et al, 2000, USA	120 patients for elective abdominal, vascular or cardiothoracic surgery. Patients randomised into 4 groups: Gp1: RIJV using landmark approach (n=30) Gp 2: RIJV using 2D ultrasound (n=30) Gp 3: LIJV using landmark approach (n=30) Gp 4: LIJV using ultrasound (n=30) Hewlett Packard 2D ultrasound probe	Prospective randomised controlled trial (level 3b)	Incidence of complications (combined arterial puncture and haematoma) in RIJV cannulation in adults	Overall 18/120 (15%) complications Incidence higher in LIJV cannulation (12/60 = 20%) versus RIJV cannulation (6/60 = 10%) P< 0.05 Complications less in Ultrasound groups (gps 2 and 4) than in controls (gps 1 and 3) Gp 1: (4 out of 30 = 13.3%) Gp2: (2 out of 30 = 6.7%) Gp 3: (8 out of 30 = 26.7%) Gp 4: (4 out of 30 = 13.3%)	End point stated as time to guidewire insertion into internal jugular vein Details of this mixed patient groups not reported - unsure how many of the patients had cardiac surgery Landmark entry point given as apex of SCM triangle No sample size calculation performed Sample size may be too small to show any striking differences Demographics not reported Level of expertise of anaesthetists stated (more than 60 IJV catheter placements)
			Number of attempts to guidewire insertion	More attempts needed for LIJV than RIJV Gp 1: (2.1 +/- 0.9) Gp 2: (1.5 +/- 2.0) Gp 3: (3.5+/-1.3) Gp4: (2.3+/- 0.7)
			Failed guidewire placement	Gp 1: (1 out of 30 = 3.3%) Gp 2: (1 out of 30= 3.3%) Gp 3: (4 out of 30 = 13.3%) Gp 4: (2 out of 30 = 6.7%)
			Time to guidewire insertion	Cannulation took longer with LIJV versus RIJV Gp 1: (137+/- 139 secs) Gp 2: (58+/- 71 secs) Gp 3: (247 +/- 176 secs) Gp 4: (138 +/- 142 secs)
Vucevic et al, 1994, UK	40 patients for cardiac surgery or ICU needing central vein cannulation. Ultrasound group: SMART needle doppler ultrasound probe (probe within needle) Patients grouped into perceived easy and difficult cannulation and then randomised to SMART needle or landmark technique 4 groups: Easy control (n=10) Easy SMART (n=10) Difficult control (n=10) Difficult SMART (n=10) Performed by 2 senior anaesthetists	Prospective randomised controlled trial (level 3b)	Incidence of complications	Combined SMART needle = 1 out of 20 carotid punctures (5%) Combined Control Group =1 out of 20 carotid punctures (5%)	Doppler guided needle device used End point stated as time to guidewire insertion into internal jugular vein Landmark entry point was lateral to carotid artery pulsation at level of cricoid cartilage Randomisation technique not described Small sample size No statistical analysis reported No sample size calculation performed
			Time to cannulation	Ultrasound Group: Easy: 91.8 secs (range 23 – 195) Difficult: 167.6 secs (range 21 – 420) Control Group: Easy 59.2 secs (range 15 – 194) Difficult 322.6 secs (range 18 –660)
			Attempts at cannulation	Ultrasound Groups: Easy: 11/10 Difficult: 23/10 Control Groups: Easy: 10/10 Difficult: 31/10
Gratz et al, 1994, USA	41 patients for cardiothoracic or major vascular surgery needing internal jugular vein cannulation Ultrasound guided Group: Central line placement with SMART needle Doppler guided needle device	Prospective randomised controlled trial (level 3b)	Incidence of complications	Ultrasound group = 0 out of 20 carotid punctures Control Group = 0 out of 20 carotid punctures	End point stated as time to cannulation of internal jugular vein with a catheter Not stated whether RIJV or LIJV cannulated Doppler guided needle device used to cannulate Landmark entry point given as apex of SCM
			Time to cannulation	Ultrasound group 109 secs (range 6 – 470 secs) Control group 226 (range 5 - 1200 secs)
			Attempts at cannulation	Ultrasound group: Successful 1st attempt cannulations = 17 out of 20 (85%) Mean number of attempts = 1.35 attempts Range 1-4 attempts Control group: Successful 1st attempt cannulations = 11 out of 20 (55%) Mean number of attempts = 2.8 attempts Range 1-10 attempts
Troianos et al, 1991, USA	160 Cardiothoracic Surgical Patients needing RIJ cannulation Ultrasound guided group: central line placement sing site rite 2D ultrasound Control Group: cannulation using landmark technique only	Prospective randomised control trial (level 2b)	Incidence of complications	Ultrasound Group: 1 out of 77 carotid punctures (1.4%) Control Group: 7 out of 83 punctures (8.4%) p=0.09	End point sated as the time of entry of needle into RIJ Landmark entry point given as apex of SCM triangle No patient demographics provided Uncertain as to whether patient population had thoracic patients included Method of randomisation not given Clinical experience of person cannulating vein not given Should have used Fischer's exact test for statistical test of carotid punctures
			Time to cannulation	Ultrasound Group: Mean 61 seconds Range 15 –180 seconds Control Group: Mean 117 seconds Range 8 - 400 seconds
			Attempts at cannulation	Ultrasound Group: Successful 1st attempt cannulation = 56 out of 77 (73%) Mean number of attempts = 1.4 attempts Range 1 – 4 attempts Control Group: Successful 1st attempt cannulation = 45 out of 83 (54%) Mean number of attempts= 2.8 attempts Range= 1-15 attempts
NICE guidelines, 2002, UK	Systematic review of the literature 20 RCTs evaluating ultrasound guidance for central line placement found Only 4 were in the adult cardiac setting -(These are the 4 identified above) 4 studies were clearly performed by non-anaesthetists	Systematic review and meta-analysis (level 1a)	Recommendations	Use of 2D USS should be considered in most clinical situations where a central line is necessary electively or in an emergency	Grades of recommendations not provided Of note no anaesthetists were on the guideline development group
			Meta-analysis of relative risks of various clinical outcome measures	Number of failed catheter placements RR: 0.16 (0.09 – 0.3) Number complications odds: 0.36 (0.17 – 0.36) risk of failure: RR 0.59 (0.39 – 0.88) Number of fewer attempts RR: 1.62 (2.57 – 0.67) Number of seconds saved 76 (96 -63) secs Number of arterial punctures saved: 90 per 1000 patients
			Cost effectiveness	Extra cost likely to be about £10 per patient, although the machine costs £7000 – 15000 initially

Comment(s)

The NICE guidelines together with their meta-analysis reported in the BMJ recommend the routine usage of a 2-D ultrasound technique for the insertion of all central lines, including those in elective settings. However these recommendations were based on papers from a very wide range of specialties, including studies in emergency medicine, Intensive care, Oncology, cardiac arrest and paediatrics. In addition it is interesting that no anaesthetists were present at the guideline development stage for these recommendations. It therefore seems that these recommendations do not seem to be based on patients similar to those in the elective cardiac surgical setting. We identified 4 papers that studied cardiac surgical patients. These were all identified by the NICE systematic review. Sulek et al studied 120 patients for elective abdominal, vascular or cardiothoracic surgery using a 2-D ultrasound technique versus a landmark approach (apex of the sternocleidomastoid triangle). They found that complications were about 50% less in the ultrasound group compared to the landmark approach, and the time to placement of the guidewire with ultrasound guidance was also less, but this was not clinically nor statistically significant. However, details of the patient population was not provided and we are unsure how many of the patients had cardiac surgery or how difficult the central line placements were perceived to be. Vucevic et al studied 40 patients (divided into predicted easy or difficult cannulation) needing central vein cannulation for cardiac surgery or in ITU using the SMART needle doppler ultrasound versus a landmark approach (lateral to carotid pulse at the level of the cricoid cartilage). The SMART needle technique utilises a small ultrasonic probe sited within the needle, which detects the motion of blood within the vessel as an audible signal. The needle is advanced in the direction of the loudest venous sound intensity (blowing low frequency venous sound as compared to the pulsatile high-pitched arterial sound) until it enters the vein. Vucevic found that the combined incidence of complications were the same in both methods (1 out of 20) and the time to placement of the cannula into the vein was greater in the difficult group (322 seconds versus 167 seconds) using the landmark approach, but this was not clinically significant. Gratz et al studied 41 patients for cardiothoracic or major vascular surgery needing central venous cannulation using the SMART needle doppler ultrasound versus the landmark approach (apex of SCM triangle). He found that there was no difference in complications between the ultrasound and landmark approach (0 out of 20) and the time to cannulations was longer in the landmark group than in the ultrasound group (226 seconds versus 109 seconds) but this was not clinically significant. However, we were unable to determine how many patients were for cardiac surgery, whether these patients were elective or urgent and what was the experience or seniority of the anaesthetist performing the procedure. Troianos et al studied a much larger group of 160 patients for cardiothoracic surgery using a 2-D ultrasound probe versus landmark approach (apex of SCM triangle). He found that the incidence of complications was higher in the landmark group (8.4%) versus the ultrasound group (1.4%) but this was not statistically significant (p < 0.09). The time to cannulation was also greater in the landmark group (61 seconds) versus the ultrasound group (117 seconds) but this was not clinically or statistically significant. Unfortunately, we are not provided with details of the number of patients who had thoracic surgery in that group, the patient demographics nor the clinical experience of the person cannulating the vein. Thus while these studies appear to show some benefit in using the 2-D ultrasound probe in predicted difficult insertions to reduce the incidence of complications and time to placement of the central line, there is insufficient evidence to recommend that the probe be used for all cases of central line insertion, especially if taking into account the expertise and skill of the operator.

Clinical Bottom Line

In patients with a potentially difficult central line insertion, the Ultrasound technique reduces complications and time to insertion. However in those patients where no difficulty is predicted, there is no evidence that the ultrasound technique confers any advantage.

References

1. Hind D, Calvert N, McWilliams R, Davidson A, Paisley S, Beverley C, Thomas S. Ultrasonic locating devices for central venous cannulation: Meta-analysis. BMJ 2003;327(7411):361-364.
2. Sulek CA, Blas ML, Lobato EB. A randomized study of left versus right internal jugular vein cannulation in adults. Journal of Clinical Anesthesia 2000;12(2):142-5.
3. Vucevic M, Tehan B, Gamlin F, Berridge JC, Boylan M. The SMART needle. A new Doppler ultrasound-guided vascular access needle. Anaesthesia 1994;49(10):889-91.
4. Gratz I, Afshar M, Kidwell P, Weiman DS, Shariff HM. Doppler-guided cannulation of the internal jugular vein: a prospective, randomized trial. J Clin Monitoring 1994;10:185-8.
5. Troianos CA, Jobes DR, Ellison N. Ultrasound-guided cannulation of the internal jugular vein. A prospective, randomized study. Anesthesia & Analgesia 1991;72(6):823-6.
6. National Institute for Clinical Excellence. Guidance on the use of ultrasound locating devices for placing central venous catheters. NICE Technical report number 49. September 2002.