Ultrasound in Central Venous Cannulation
More than 5 million central venous catheters (CVCs) are placed yearly [1] for a multitude of indications, and this number is expected to increase due to an aging population. With demands to prevent medical errors coupled with the improvement of portable ultrasound technology, the role of ultrasound guidance in central vein catheterization grows increasingly important. This article is not another meta-analysis, but looks at the significant history behind ultrasound’s prominence in guiding central venous access and provides the rationale as to why ultrasound should be adopted in current anesthesia practice. The second part of this article reviews the technical skills with which to achieve success with ultrasound in central venous access.
By the 1980s, ultrasound in central venous catheterization had already been well described [2,3]. Many of the initial studies comparing ultrasound to surface landmarks had differing methodologies that made generalization difficult. For example, the ultrasound-guided dynamic technique is likely superior to the static technique, that is, real-time guidance (visualization and identification of the relevant anatomic structures, tracking the progression of the needle, and confirmation of the guidewire within the target central vein) as compared with ultrasound being used for the sole purpose of anatomic identification and subsequent surface marking before needle insertion [4,5], yet each is considered ultrasound-assisted. In addition, there are differences in what constitutes a failure (number of attempts, an absolute time in procedure, or carotid puncture). Another confounding variation is which surface landmark approach ultrasound is to be judged against, as there are numerous landmark-based approaches. Even in this decade there are still studies [6,7] comparing the “superiority” of one landmark-based approach over another, yet it is uncertain how much operator experience and frequency of CVC placements contributes to success and complication.
Two meta-analyses are worth reviewing, as they are the most frequently cited, and often the basis for recommendations for the use of ultrasound guidance in central venous cannulation. Randolph and colleagues [8] identified 8 randomized controlled trials comparing ultrasound guidance versus traditional surface landmark-based techniques. There was a decrease in the number of placement failures in both internal jugular (relative risk [RR] 0.26; 95% confidence interval [CI] 0.11–0.58) and subclavian veins (RR 0.11; 95% CI 0.02–0.56) when ultrasound was used. The number of attempts before success also decreased with the use of ultrasound (RR 0.60; 95% CI 0.45–0.79). Of note, operator experience was not reported in many of the studies, and it was uncertain if there was any prior formal training or exposure to ultrasound concepts and equipment. In addition, the use of color flow Doppler and 2-dimensional (2D) ultrasound were mixed and considered to be equivalent.
In 2003, Hind and colleagues [9] performed a larger meta-analysis consisting of 18 trials (1646 patients) comparing 2D ultrasound with landmark-based techniques, specifically looking at failed catheter placement, catheter placement complications, number of attempts, and time to cannulation. Their results were more conclusive, as ultrasound decreased failed catheter placements both for internal jugular (RR 0.36; 95% CI 0.11–1.19) and for subclavian vein approaches (RR 0.09; 95% CI 0.02–0.38). As most of the data were based on the internal jugular approach, the time to catheterization with ultrasound was faster, 180 seconds, versus the landmark-based technique of 192 seconds (P <.0001). Though statistically significant, 12 seconds may not confer great time-saving advantage, as it may take at least that amount of time to locate the ultrasound machine. Fig. 1 shows a forest plot of the randomized controlled studies comparing failed catheter placement of 2D ultrasound versus the landmark technique at the internal jugular vein location.
Although not as extensively studied, the reviewers also supported ultrasound’s role in the subclavian and femoral venous approaches despite small sample sizes of 20 to 25 patients in each study arm. Similarly, ultrasound’s use in pediatrics was advocated in the internal jugular region based on data compiled from 3 studies with about 80 patients in each division. Newer small studies and reports have continued to show favorable outcomes with ultrasound-guided femoral [10] and subclavian vein [11] catheterizations.
These often quoted meta-analyses along with growing public and medical awareness of central line complications prompted the Agency for Healthcare Research and Quality (AHRQ) of the US Department of Health & Human Services in 2001 to issue the following statement after their own investigation: “Real-time US guidance for CVC insertion…improves catheter insertion success rates, reduces the number of venipuncture attempts prior to successful placement, and reduces the number of complications associated with catheter insertion” [12]. The British Committee for Standards in Haematology issued their guidelines [13] for central venous access insertions in 2007, recommending the use of ultrasound “for all routes of central venous catheterization,” and by 2008 the American College of Surgeons published guidelines that support the uniform use of real-time ultrasound guidance for the placement of CVCs in all patients [14]. A recent editorial has advocated that ultrasound for internal jugular central venous catheterization in the critical care setting should become the standard of care [15].
As practicing anesthesiologists, is there still a role for ultrasound? After all, the statements above did not deem it absolutely necessary for anesthesiologists to use this technology as the AHRQ wrote, “as experienced anesthesiologists can continue to place most CVCs without US guidance.” In fact, surveys on use suggest ultrasound use is actually low among anesthesiologists. Bailey and colleagues [16] sent out an electronic survey to all members of the Society of Cardiovascular Anesthesiologists and with a response rate of 35% showed that two-thirds never or almost never use ultrasound guidance versus 15% who always or almost always use it. While ultrasound availability was absent in 18% of respondents, the most common reason (46%) for not using ultrasound is the perception that it is not needed, despite almost 75% of the respondents having experienced a previous carotid puncture, 16.7% a pneumothorax, and 1.1% a stroke due to CVC placement.
Other investigators have also found that ultrasound use is low despite having access to an ultrasound machine. After the United Kingdom’s National Institute for Clinical Excellence recommended 2D ultrasound guidance, a survey of pediatric anesthetists in the United Kingdom also found lower than anticipated use. With a response rate of 63% and availability of ultrasound in 82% of the workplaces and 74% receiving 2D ultrasound training, only 26% of anesthetists with access to ultrasound used it on elective cases [17]. A separate postal survey in the United Kingdom found similar responses, with only 39% of pediatric anesthesiologists routinely using ultrasound guidance. The majority used either ultrasound or a landmark-based technique depending on the clinical circumstance, despite widespread access to and education in ultrasound [18]. However, the adaptation of this technology for central venous cannulation is taking place. A large German survey of anesthesia departments in 2007 revealed ultrasound use at 40%, much improved from 2003 when it was at 19% [19]. Thus the majority of CVCs are still placed via the landmark-based technique, and studies comparing landmark-based approaches continue to be published [7].
Ultrasound use in specialties outside of radiology is already widespread, including cardiology (transesophageal or transthoracic echocardiography), regional anesthesia for peripheral nerve blocks [20], and critical care and emergency medicine [21]. For example, transesophageal echocardiography has supplanted pulmonary artery catheterization for hemodynamic assessment, ultrasound-guided pericardiocentesis is superior to blind, electrocardiographic-guided, or fluoroscopically-guided techniques, and the focused abdominal sonogram for trauma (FAST) is often considered superior to diagnostic peritoneal lavage. Ultrasound in vascular access may be more readily embraced by emergency and internal medicine than by anesthesia. Nomura and colleagues [22] revealed a remarkable 90% ultrasound use among internal medicine and emergency medicine residents at a tertiary hospital, with most (88%) believing it to be easier than a landmark-based technique and 74% of respondents believing it should be used for all CVC placements. The investigators’ rationale for such high ultrasound use was attributed to increased adoption by the physicians-in-training versus those who have finished postgraduate training and, in addition, the presence of an emergency medicine ultrasound fellowship program.
Like many other great “discoveries” in medicine, it often takes many years before evidence-based practices are integrated at an institutional level. An editorial response to one of the surveys showing low ultrasound use suggested the implementation of hospital-wide clinical protocols for the routine use of ultrasound, as was successfully employed in the Veteran Affairs system [23]. Fig. 2 depicts the major steps in integrating evidence-based medicine in daily practice. The 2 arrows point to the adaptation points where most of medicine stands on ultrasound’s role in central venous access. Minimal new evidence from prospective randomized studies has been published on ultrasound for CVC placement in the last few years. Many local and national levels are developing the evidence-based clinical policies, on the plethora of existing data, and applying ultrasound in routine CVC placements. Of note, the American Society of Anesthesiologists (ASA) (Rupp, personal communication, 2010), National Institutes of Health, and the Cochrane Collaboration [24] are expected to release guidelines on ultrasound’s importance in central venous catheterization. It is very likely that their statements will further accelerate ultrasound adoption.
What really is, then, the incidence of complications of central venous catheterization? After all, complications [25] are often unnoticed unless severe, and more often than not they are underreported or not surveyed. Prior to ultrasound, the ASA Closed Claims project in 1996 gave some insight as to what constitutes complications [26] (Table 1). By self-reporting without an actual denominator, the registry helped promote awareness but did not create urgency despite the report of deaths associated with central venous cannulation, because these numbers appear small compared with the perceived number of central venous lines placed. In one of the earliest articles comparing ultrasound to landmark anatomy [27], carotid artery puncture occurred in 8.3% of patients and brachial plexus irritation occurred in 1.7% of patients in the traditional landmark-based technique versus 1.7% and 0.4% with ultrasound, respectively, in the “experienced” hands of cardiology fellows and attending physicians.
Table 1 The ASA Closed Claims project provides some of the dangers associated with central venous catheterization
| Complication | Total | Fatalities |
|---|---|---|
| Cardiac tamponade | 11 | 10 |
| Wire or catheter embolism | 12 | 0 |
| Vascular injuries (nonpulmonary artery): | 13 | 5 |
| Hemothonax | 6 | 4 |
| Hydrothonax | 3 | 1 |
| Carotid artery injury | 3 | 0 |
| Subclavian artery aneurysm | 1 | 0 |
| Pulmonary artery rupture | 2 | 2 |
| Pneumothorax | 7 | 1 |
| Air embolism | 2 | 2 |
| Fluid extravasation in neck | 1 | 0 |
| Total | 48 | 20 |
Data from Bowdle T. Central line complications from the ASA Closed Claims Project. ASA Newsl 1996;60:22.
Furthermore, being able to visualize the anatomic structures has other advantages. Asouhidou and colleagues [28] performed cadaveric dissections in 93 patients, and showed 3 internal jugular veins less than 6 mm in diameter with corresponding enlarged ipsilateral external jugular veins. These small internal jugular veins had no evidence of thrombosis, stenosis, and recent or previous cannulation, and traveled the usual anatomic course. This 3% of “absent” internal jugular veins was similar (2.5%) to that found in the 1991 ultrasound survey by Denys and Uretsky [29] of 200 patients in the intensive care unit and cardiac catheterization laboratory. With that discovery and an additional 2% of patients demonstrating the internal jugular vein located medial to the carotid artery, Denys and Uretsky recommended ultrasound examination before any attempted cannulation. Other investigators [30,31] have also confirmed the existence of smaller internal jugular veins, and consider this a “powerful predictor” for prolonged procedure time and carotid puncture. Mey and colleagues [30] concluded that an internal jugular vein smaller than or equal to 7 mm, when normally averaging 1.0 cm (range 0.46–3.6 cm), is an independent predictor of catheterization failure (P = .001).
Ultrasound also has benefit in reducing complications in children. In a 4-year retrospective study, Tercan and colleagues [32] compared pediatric (average age 3.3 years) with adult (56.3 years) patients and found similarly low complication rates (2.3% vs 2.4%) when ultrasound was used. No major mechanical complications such as pneumothorax or hemothorax were noted in 859 adult and 247 pediatric catheterization attempts.
Normal anatomic variations, probe position, and head rotation all influence the relationship of the internal jugular vein and carotid artery. For example, internal jugular vein duplication is estimated at 4 in 1000 patients [33]. As expected, the internal jugular vein is not always immediately anterior and lateral to the carotid artery as previously described [2]. Lin and colleagues [34] described variations in the position of the internal jugular vein in as many as 17.3% in 104 consecutive uremic patients for dialysis catheters. Another similar study in 450 nonuremic patients also noted that the “classic” anterior lateral position is found only in 79.3% on the right side and 83.5% on the left [35].
Head rotation (as is commonly used to help facilitate the classic surface landmark-based techniques in internal jugular venous cannulation) may potentially increase the risk of carotid injury. In a large ultrasound study, Troianos and colleagues [36] recorded the images of 1136 patients with normal surface anatomy for central venous placement to determine the relationship of the right internal jugular vein relative to the carotid artery. All patients were in the supine position without a pillow, with their heads rotated as far to the left as was comfortable. After independent scoring into 5 segments (completely lateral, up to 25% overlap, 25%–50%, 50%–75%, and >75% overlap) by 3 investigators, they concluded that the internal jugular vein overlies the carotid artery by 75% or more in 54% of all patients. Arai and colleagues [37] demonstrated in infants and children that head rotation to 45° as opposed to 0° at the cricoid level increases the overlap of the internal jugular vein to the carotid artery. Lieberman and colleagues [38] further investigated the influence of head rotation by comparing rotations at 0°, 15°, 30°, 45°, and 60° left of the midline in simulated catheterization of volunteer adults. This group found that the occurrence of carotid artery contact with the internal jugular vein is increased with excessive head rotation. Recommendations for optimal rotation in patients with body mass index (BMI) greater than 25 kg/m2 is 30° and 45° to 60° for patients with BMI <25 kg/m2. Fujiki and colleagues [39] also found a much higher degree of vessel overlap in obese patients when the head was rotated away from the neutral position. Thus, ultrasound may allow the operator to decide on the optimal approach, and the optimal neck positioning may indeed be neutral, contrary to prior practices for landmark-based approaches.
Of course, this would lead to the question, how often does carotid puncture occur and does it matter? The likely answer is it happens more frequently than reported. The Mayo Clinic answered this with a prospective study of 1011 consecutive cardiothoracic and vascular surgery patients, and found the incidence of carotid puncture with a “finder” needle at 9.3% with a landmark-based technique for internal jugular vein cannulation [40]. Damen and Bolton [41], in another large prospective study of 1400 patients scheduled for cardiac surgery, found an incidence of 4.8%, and in the pediatric population the reported incidence is 7% to 8% [42,43].
The adverse consequences of arterial punctures with a “finder” needle are likely less than that of dilation and placement of a large-bore catheter. Because of successful early detection of arterial puncture by pressure transduction or manometry in the Mayo study [40]
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
