Presented at the Society for Academic Emergency Medicine Annual Meeting in Washington, DC, May 2008. This study was supported by the Washington University School of Medicine CTSA Grant (UL1 RR024992) and the Washington University School of Medicine’s Clinical Research Training Center (KL2 RR024994).
A Descriptive Comparison of Ultrasound-guided Central Venous Cannulation of the Internal Jugular Vein to Landmark-based Subclavian Vein Cannulation
Article first published online: 2 APR 2010
© 2010 by the Society for Academic Emergency Medicine
Academic Emergency Medicine
Volume 17, Issue 4, pages 416–422, April 2010
How to Cite
Theodoro, D., Bausano, B., Lewis, L., Evanoff, B. and Kollef, M. (2010), A Descriptive Comparison of Ultrasound-guided Central Venous Cannulation of the Internal Jugular Vein to Landmark-based Subclavian Vein Cannulation. Academic Emergency Medicine, 17: 416–422. doi: 10.1111/j.1553-2712.2010.00703.x
- Issue published online: 2 APR 2010
- Article first published online: 2 APR 2010
- Received August 20, 2009; revision received October 15, 2009; accepted October 19, 2009.
- central venous;
- emergency medicine
Objectives: The safest site for central venous cannulation (CVC) remains debated. Many emergency physicians (EPs) advocate the ultrasound-guided internal jugular (USIJ) approach because of data supporting its efficiency. However, a number of physicians prefer, and are most comfortable with, the subclavian (SC) vein approach. The purpose of this study was to describe adverse event rates among operators using the USIJ approach, and the landmark SC vein approach without US.
Methods: This was a prospective observational trial of patients undergoing CVC of the SC or internal jugular veins in the emergency department (ED). Physicians performing the procedures did not undergo standardized training in either technique. The primary outcome was a composite of adverse events defined as hematoma, arterial cannulation, pneumothorax, and failure to cannulate. Physicians recorded the anatomical site of cannulation, US assistance, indications, and acute complications. Variables of interest were collected from the pharmacy and ED record. Physician experience was based on a self-reported survey. The authors followed outcomes of central line insertion until device removal or patient discharge.
Results: Physicians attempted 236 USIJ and 132 SC cannulations on 333 patients. The overall adverse event rate was 22% with failure to cannulate being the most common. Adverse events occurred in 19% of USIJ attempts, compared to 29% of non–US-guided SC attempts. Among highly experienced operators, CVCs placed at the SC site resulted in more adverse events than those performed using USIJ (relative risk [RR] = 1.89, 95% confidence interval [CI] = 1.05 to 3.39).
Conclusions: While limited by observational design, our results suggest that the USIJ technique may result in fewer adverse events compared to the landmark SC approach.
ACADEMIC EMERGENCY MEDICINE 2010; 17:416–422 © 2010 by the Society for Academic Emergency Medicine
Physicians in the United States insert over 5 million central venous catheters (CVC) annually.1,2 In studies that examine acute mechanical complications, no anatomical site proves more advantageous than the other.3,4 Two emergency department (ED) randomized clinical trials demonstrate that ultrasound (US) guidance decreases the mean number of needle sticks required to successfully cannulate the internal jugular vein and modestly reduces adverse events.5,6 However, these studies involved groups of physicians with an interest or special expertise in US-guided procedures. Moreover, hospital surveys indicate that physicians may favor the subclavian (SC) approach over other sites.7,8 Therefore, in this study, we sought to compare adverse events among patients who underwent US-guided cannulation of the internal jugular vein (USIJ) to that in patients undergoing the non–US-guided SC approach in a setting where physicians without focused training in either technique performed the procedure.
We assembled a prospective cohort of adult patients who underwent CVC in the ED. The institutional human research protection office approved the study and waived informed consent.
Study Setting and Population
We conducted the study in an academic, urban, 63-bed ED with a volume of 79,000 visits per year. The emergency medicine training program includes 48 residents total (12 per year). During the year the study took place, 42 full-time emergency physicians (EPs) staffed the ED.
We collected a convenience sample of consecutive CVC attempts in our ED. The study was conducted from March 20, 2007, to March 17, 2008. We included any patient who underwent cannulation of either the internal jugular or SC vein, as determined by the treating EP caring for the patient. We excluded patients with a concurrent indwelling CVC, suspected traumatic injury of the vessels in the neck or thorax, or overlying dermatologic process that precluded insertion at either site and those who underwent CVC at another anatomical site.
To capture consecutive central line attempts, we created a standardized electronic procedure note for the ED’s electronic medical record system. Physicians working in the ED noted the indication for the procedure, the anatomical site of the procedure, if US guidance was employed, and any adverse event that may have occurred during the procedure. Any patient who required central venous access due to lack of peripheral access was defined as “poor access.” Any patient who required central venous access for resuscitation or hemodynamic monitoring (as in cases of presumed sepsis) was defined as “septic/hypovolemic or hypotensive.” If the central line was inserted while the patient was in cardiopulmonary arrest (from either traumatic or medical reasons) the indication for cannulation was considered as “code line.” USIJ was defined as the real-time use of an US probe to direct the locator needle toward the internal jugular vein. We excluded central lines placed via the static approach (using the US probe to site the vein but not actively cannulate it). SC vein cannulation was defined as any infraclavicular attempt to cannulate the SC vein, without US guidance to locate the vessel or direct the needle. We specifically defined an “attempt” as a single operator’s performance of the procedure from beginning to completion at one anatomic site. If a second operator attempted the procedure at the same site, this was considered a new attempt. Because the purpose of our study was to examine adverse events and not efficiency outcomes, we did not include the number of “needle sticks” or time required for either procedure in this analysis. To improve the precision of our data collection instrument, we made two semiannual presentations regarding our definitions of each variable. We did not modify the data collection instrument during the study.
Demographic characteristics and vital signs were collected from the patients’ ED charts. Body mass index (BMI) was calculated according the patient’s self-reported height and weight at the time of admission to the hospital and categorized according to World Health Organization BMI classifications. If no weight was recorded, the weight in the ED chart recorded by the ED nurse caring for the patient was noted. The height at the nearest prior admission or from the patient’s identification card copied into the chart was recorded if no other record existed. The International Normalized Ratio (INR) was obtained from the ED chart. Any patient with systolic blood pressure (sBP) of ≤90 mm Hg at any time during the ED stay was considered hypotensive. Patients who “required pressors” were identified by review of pharmacy records to determine whether the patient received vasopressors (epinephrine, norepinephrine, or dopamine) or inotropes (dobutamine) at any time during the ED stay. We defined “mortality” as those patients who did not survive the ED to hospital admission and those who expired during their hospitalization.
All of the operators participating in the study were physicians. The operators underwent training for each technique under the “apprenticeship” model, meaning they gained central line experience while simultaneously clinically caring for patients. There were two additional lectures provided for US-guided techniques, but there were no separate simulation or “hands-on” sessions offered. All EPs were asked to fill out a survey indicating their overall experience with CVC. The survey asked the physicians how many central lines, irrespective of site or technique, they had placed in increments of 10 to a maximum of 50 or above. We identified no prior validated definition of “experienced operator,” and all operators indicated they had placed more than 20 central lines throughout their career, a threshold used in prior studies.5 We defined “experienced operator” as one who had performed more than 50 cannulations (not site specific) throughout their career since our survey results indicated that a majority of the operators had attained this level of experience. Level of experience was missing for 16 attempts (5%). All were from physicians with four or more years of training.
Our primary outcome of interest was a composite of acute adverse events. Borrowing from prior definitions, we considered an adverse event to occur when a cannulation attempt resulted in an accepted complication of, or failure in, its technical performance.9–12 We therefore defined an acute adverse event as a composite of the following outcomes: arterial puncture, pneumothorax, hematoma, extravascular or misplacement, and failed attempt at cannulation. An arterial puncture was judged to have occurred when physicians aspirated pulsatile arterial (bright red) blood into an 18-gauge locator syringe. The presence of a pneumothorax was determined using plain films of the chest (either standard posteroanterior or portable anteroposterior) as read by a radiologist blinded to the technique used for central line cannulation. Any ipsilateral pneumothorax identified within 24 hours of line placement was considered a complication of the procedure unless the patient underwent another procedure that could also cause a pneumothorax on the side in question. Physicians or nurses documented a hematoma if an area of bruising or swelling greater than 3 cm was noted around the insertion site. Placement of the central line outside of the vascular lumen was defined by failure to aspirate blood through any of the ports or a visible extravasation of intravenous fluid in the soft tissues (a “blown” line appearance). A failed attempt at cannulation was specifically defined as the inability to complete the intended procedure by the primary operator for lack of successfully cannulating the intended vessel. Any purported puncture of the vein that resulted in the inability to pass a guide wire, and therefore complete the procedure, was considered a failed attempt. We did not attempt to determine the number of needle passes for this analysis.
In the event that more than one adverse outcome occurred, we recorded the complication that required a therapeutic intervention as the primary adverse outcome. We considered pneumothorax, failed cannulation, and misplacement as primary adverse events if they occurred simultaneously with arterial puncture and hematoma. However, any occurrence of any complication was considered an “adverse event.”
We followed subjects until there was radiologic or medical record evidence that the ED central line was removed or replaced, or to hospital discharge, to examine whether seemingly insignificant acute complications resulted in any therapeutic intervention or delayed complications. A chart reviewer (SP and MP), blinded to the site and technique of insertion, reviewed daily nursing notes for the mention of “hematoma” or “transfusion.” Patient discharge summaries were examined for evidence of a central line related adverse event. Any mention of “central line complication” or “cannulation device complication” triggered an audit of inpatient nursing and operative records. Operative records were reviewed for any mention of a vascular procedure related to central line complications. In addition we reviewed the annual records of our quality improvement process for any cases referred to the committee that involved an ED central line. For the purposes of this study, catheter-related infections or thrombotic events were not recorded.
Results were inserted into an SAS (Version 9.2, SAS Institute, Cary, NC) database. We performed bivariate analysis for variables of interest using chi-square tests and report the 95% confidence intervals (CI). Power calculations were performed using an assumption of a 15% acute complication rate using the external landmark technique and a 5% acute complication rate using the US-guided technique. Using these figures and accounting for dropouts, we determined that it would require 128 patients in each group to have a power of 0.80 to detect this difference at the 0.05 level of significance.
Physicians reported a total of 729 CVC attempts during a 1-year period from March 20, 2007, to March 17, 2008 (Figure 1). Almost one-third (32.8%) were inserted into the femoral vein. We excluded those central lines inserted into the internal jugular vein without the use of US, and those inserted into the SC vein under US guidance, as well as all central lines inserted by physicians who did not meet our minimum requirement of 20 central lines during their career (16.2%). Our final sample consisted of 333 patients who underwent cannulation by 63 physicians. Physicians indicated that 254 (69%) central lines were inserted for shock, 97 (26%) were inserted for lack of peripheral access, and 17 (5%) were inserted for cardiopulmonary resuscitation (Advanced Cardiac Life Support or Advanced Trauma Life Support protocols). A total of 32 physicians (51%) contributed one attempt to each group, and 17 physicians (27%) contributed two or more attempts to each group. Physicians indicating that they had placed more than 50 central lines throughout their career made 275 attempts (76%). Physicians in their first, second, and third years of training contributed 35 attempts (10%), 61 attempts (17%), and 104 attempts (28%), respectively. Attending physicians and those in their fourth year of training contributed 168 attempts (46%).
Overall there were 82 adverse events (22%, 95% CI = 18.2% to 30.0%). Of the 236 central lines inserted using the USIJ technique, 44 (19%) resulted in an adverse event. Of the 132 attempts made at the SC, 38 (29%) resulted in an adverse event (Table 1). There was no statistically significant difference between USIJ and SC attempts among any clinical or demographic characteristics, except age and mortality. Physicians who met our definition of “experienced operator” used the SC more often than the USIJ technique.
|US-guided Internal Jugular Approach (n = 236), n (%)||Landmark SC Approach (n = 132), n (%)||%Δ||95% CI|
|Age ≥ 65 years||109 (46.1)||47 (35.6)||10.6||0.6 to 20.5|
|Male sex||107 (45.3)||73 (55.3)||–10.0||–20.0 to 0.6|
|African American||139 (58.9)||81 (61.3)||–2.4||–8.4 to 13.0|
|COPD/asthma||35 (14.8)||20 (15.1)||–7.9||–17.1 to 0.6|
|Hypotensive in the ED||155 (65.6)||97 (73.4)||–7.8||–17.4 to 2.5|
|Requiring pressor*||87 (36.8)||57 (43.1)||–6.3||–17.1 to 4.4|
|BMI > 30||76 (32.2)||32 (24.2)||7.9||–2.2 to 17.3|
|INR > 2.0||41 (17.4)||15 (11.4)||6.0||–2.3 to 13.3|
|ACLS/ATLS||8 (3.4)||9 (6.8)||–3.4||–9.8 to 1.4|
|Mortality||40 (16.9)||37 (28.0)||–11.0||–20.8 to –1.9|
|Inserted by operator with >50 CVCs in career||162 (72.0)||113 (89.0)||–17.0||–25.8 to –7.9|
|Adverse event||44 (18.6)||38 (28.8)||–10.2||–19.5 to –1.2|
The most common adverse event across both sites was failed attempt at cannulation. In total, 53 attempts at initial cannulation failed. In the SC group, 30 (23%) cannulation failures and two (2%) pneumothoraces occurred. Both pneumothoraces required thoracostomy. SC attempts resulted in two (2%) hematomas and three (2%) arterial punctures, and one (1%) catheter was misplaced in the soft tissues. The misplaced catheter resulted in no therapeutic intervention after removal. The USIJ technique failed to result in cannulation in 23 (10%) attempts and zero pneumothoraces occurred. There were 17 (7%) hematomas noted, four (2%) inadvertent arterial punctures, and zero misplacements.
We found no instance where a delayed complication occurred or where a seemingly insignificant complication in the ED required direct therapeutic intervention upon follow-up in either group. One USIJ resulted in a delayed hematoma 48 hours after the time of admission. The patient was on heparin therapy and did not undergo any further therapeutic intervention, nor was heparin therapy discontinued. The quality improvement team reported no additional adverse events.
In general, more adverse events were noted among those patients undergoing attempted cannulation of the SC compared to those undergoing USIJ cannulation, regardless of operator experience. The adverse event rate decreased with operator experience regardless of approach used (Figure 2). Cannulations performed by experienced operators were more likely to result in an adverse event at the SC site than those performed using USIJ (relative risk [RR] = 1.89, 95% CI = 1.05 to 3.39), but there was no difference among novice operators in adverse events when using the SC technique (RR = 2.2, 95% CI = 0.66 to 7.3). There was no significant difference in adverse events by postgraduate year (χ2 = 1.53, p = 0.68; Figure 3).
Although our study was observational by design and therefore limited, our findings suggest that physicians who adopt US guidance when cannulating the internal jugular in an “apprenticeship model” are less likely to encounter an adverse event compared to the landmark SC technique. Past studies that compared vascular access sites to one another did not account for the impact of US guidance. In these studies, adverse events occurred with nearly equal proportion in both groups, and no anatomic location was conclusively superior in regard to acute mechanical complications.2,4,13–20 While a randomized clinical trial would be required to recommend a change in practice, our observational study suggests a benefit to using US-guided cannulation.
Our study also found that adverse events occurred less frequently with the USIJ technique compared to the SC technique among experienced operators. Past studies suggest that US guidance conveys its greatest benefit on operators with less experience, while more experienced operators may realize less improvement in their cannulation outcomes.21 Our findings suggest that the US-guided technique confers benefit to experienced operators in the absence of rigorous standardized training in either technique. Measuring expert operator performance by strict counts may have limitations. However, accreditation boards and prior research in the area have focused on the number of procedures performed to assess technical competence.5,22–24 Our study also relied on the number of performed procedures, but it was self-reported and not site-specific. Among our most experienced operators (those who performed >50 CVC in their careers), USIJ resulted in fewer adverse events, specifically fewer cannulation failures. Although we caution the interpretation of an individual component of our composite, our study suggests that even operators with significant experience can gain benefit from US-guided procedures. Overall experience with cannulation may facilitate the acquisition of skills that pertain to US-assisted procedures.
With the exception of two pneumothoraces that required chest tubes, none of our 333 patients sustained an adverse event that required clinical interventions. These findings echo past studies suggesting that the majority of acute adverse events resulting from attempted CVC in the ED are typically “no harm events.” While this observation ignores the impact that failed access may have on resource consumption and timeliness of therapeutic interventions, it underscores the modest extent to which CVC leads to adverse events requiring intervention.
In prior studies comparing US-guided attempts to non-US-guided attempts at a single site, the decrease in adverse events that required a direct intervention was modest as well. In the study by Milling et al.,6 no attempts resulted in a pneumothorax, and in the study by Leung et al.,5 1 of 65 attempts resulted in a pneumothorax in the non-US group. Both studies demonstrated fewer hematomas and arterial punctures, but did not report any therapeutic consequences.5,6
Hematomas were more common in our USIJ group than in previous studies. This may be due to the fact that our physicians were not selected for demonstrated expertise in USIJ and may better reflect the true incidence of this complication in a “real-world” experience. Hematomas were reported significantly less often in the SC group than in the USIJ group. This could be attributed to the difficulty in identifying hematomas in the subclavicular location. The effect of this would be to bias the study toward the SC approach. However, no hematoma in either group resulted in a clinical intervention in our cohort. Whether resource utilization, timeliness of therapy, and a decrease in thoracostomies becomes sufficiently significant across larger populations to justify solely using US-assisted techniques remains a question for further investigation in larger trials.
Our cohort sustained a high failure rate utilizing US compared to prior studies in similar settings.5,6,21,25 We attributed this to a lack of hands-on training or simulator-based learning opportunities. In prior ED studies, US-guided cannulations (intervention arms) were performed by small groups of physicians with targeted training suggesting possible “operator bias” frequently noted in US research. The study by Milling et al.6 involved nine physicians, and the study by Leung et al.5 involved 13 physicians, compared to the 63 physicians involved in our cohort. The 63 physicians in our cohort did not undergo intense targeted training in the use of US assistance, but adopted the procedure in the classic apprenticeship model. Despite the lack of focused training, the intervention of US assistance resulted in some benefit. Whether the impressive decrease in US-guided central line complications noted in other studies is a factor of directed educational sessions or a trend only among physicians with highly specialized training in US remains unexplored.
The adoption of the US-guided technique came at the expense of performing fewer non–US-guided central lines. This may be partly explained by the suggestion from prior studies that physicians in training may benefit the most from US guidance.21,26 This may represent a concerning practice shift, since failed US-guided attempts may require rescue by non–US-guided techniques. Given that a majority of community EDs have limited access to US technology, training programs should be wary of overemphasizing US-guided techniques so as to not impede the development of non–US-guided techniques among their graduates.27
Our SC adverse event rate was higher than in studies conducted by experienced operators, yet lower than in some studies of inexperienced operators, a fact likely explained by the nature of emergent central lines and the acutely ill ED patient population.16,26,28,29 The mortality rate of patients undergoing the SC technique was higher than those undergoing USIJ. Our data suggested that this was due to physician preference for landmark-based approaches during “code” situations. The poor prognosis of these patients likely explains the baseline difference in mortality rates between the two central line approaches. Whether US guidance can assist placement of SC vein catheters in the ED requires further investigation.
Due to the observational design, central line attempts were not randomly assigned, increasing the likelihood that selection bias was introduced. We hypothesized that physicians would preferentially use the technique with which they were most comfortable, possibly biasing our results toward the null. Despite this, we still found a difference in adverse events, even among those considered experienced with CVC. Although we accounted for patient factors that might make venous cannulation difficult (e.g., BMI), it is possible that several unrecorded patient variables could result in a selection bias toward one technique or the other.
We relied on the “apprenticeship model” to teach both techniques. It is possible that an institutional trend developed whereby USIJ became the preferred method. This may have biased the results in favor of USIJ in that less time was devoted to acquiring skills of non-US techniques. However, a survey during the study indicated that 80% of all physicians were initially taught solely non–US-guided techniques, and 68% did not use US in almost half their central line attempts.30 However, as the study proceeded, more physicians opted for US guidance than the landmark technique.
As in prior studies, our measure of physician experience was based on survey data. The results of our survey are subject to recall bias. Furthermore, we asked physicians to indicate their general experience with CVC, not their experience with each procedure and not their experience with US. This may bias operator experience results because physicians may have reported experience solely with one technique and at one site, not both. In this case, adverse events may be explained more by experience than technique. Although we restricted our analysis to exclude “beginner”-level physicians, we were not able to confirm the level of experience of operators for each technique. However, current studies and accreditation standards currently do not define experience other than by the total number of cannulations performed during training or career.
It is likely that physician and nurse self-report led to measurement bias. Other than pneumothorax, unsuccessful cannulation, and misplacement, the outcomes studied do not have gold standards and rely on reporting by the physician or the nurse. In addition, some adverse events may be more obvious in certain anatomic areas. Furthermore, our results would be greatly biased if physicians reported their successes less commonly with their preferred technique while they reported their failures more commonly with their least preferred technique or vice versa. However, the rates of our adverse events were similar to those found in other studies, and our quality assurance committee found no instance of an unreported adverse event that took place in the ED during this study.
Our observational trial suggests that physicians without rigorous standardized training can adopt the ultrasound-guided internal jugular technique and decrease the number of acute adverse events, compared to using the landmark subclavian approach. However, our study was limited by its design, and a randomized control trial would be necessary to recommend a change in practice. Very few adverse events from central venous cannulation attempts require direct clinical intervention. Whether the benefit of ultrasound guidance translates to other outcomes, such as timeliness of therapy, resource utilization, or thrombotic or infectious complications, remains unstudied and will likely require large coordinated efforts to gain statistical power. Physicians who solely perform the landmark subclavian approach may expose their patients to a modestly increased risk of an acute adverse event, mainly the risk of failing to complete the intended procedure.
The authors acknowledge Sarah Boslaugh, PhD, and Missy Krauss, MPH, for their statistical advice for this work. We also acknowledge Max Palatnik, Connor Deal, and Shruti Patil for their contributions to this project.
- 12002 National Hospital Discharge Survey. Adv Data. 2004; 342:1–30., .
- 7Prevalence of the use of central venous access devices within and outside of the intensive care unit: results of a survey among hospitals in the Prevention Epicenter Program of the Centers for Disease Control and Prevention. Infect Control Hosp Epidemiol. 2003; 24:942–5., , , et al.
- 21Ultrasound guidance of central vein catheterization. In : MarkowitzAJ (ed). Making Health Care Safer. San Francisco, CA: University of California at San Francisco (UCSF)-Stanford University Evidence-based Practice Center, 2001, pp. 245–53..
- 22Accreditation Council for Graduate Medical Education. Emergency Medicine Guidelines, 2009. Available at: http://www.acgme.org/acWebsite/RRC_110/110_guidelines.asp. Accessed Jan 8, 2010.
- 23Development and validation of a scale for the clarity of anatomical landmarks for use in outcomes research of ultrasound guidance of central venous cannulation [abstract]. Acad Emerg Med. 2006; 13:e368., , .