| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
Department of Anesthesiology, University of Rochester, Rochester, New York
Address correspondence and reprint requests to Peter Bailey, MD, Department of Anesthesiology, Box 604, 601 Elmwood Avenue, Rochester NY, 14642. Address e-mail to peter_bailey{at}urmc.rochester.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
There are numerous approaches to central venous catheterization of the right IJ (RIJ) vein, including "posterior," "anterior," and "para-carotid" approaches (2,6–9). The "central" landmark, first described by English et al. (10,11) is commonly used. This approach includes turning the head away from the side to be catheterized, defining the apex of the triangle formed by the 2 heads of the sternocleidomastoid muscle as the sides of a triangle and the clavicle as the base, (12) and inserting a needle at approximately 30–45 degrees to the coronal plane at this apex, aiming towards the ipsilateral nipple. English et al. (10) reported a success rate of 93.5% with no significant complications. However, landmark-based approaches to central venous catheterization, compared with catheterization guided by ultrasound, have been demonstrated to be associated with more frequent complications (13–16).
Some of the anatomic reasons for the limitations of landmark-based approaches, such as variations in the size of the IJ vein or its relationship with the carotid artery, have been previously described (14,17). The purpose of this study was to evaluate the accuracy of the central landmark approach. We used the digital cursor incorporated in echocardiographic equipment to simulate a needle path based on the landmark and then determined how often the simulated needle path intersected the IJ vein and did not intersect the carotid artery. We validated this simulation approach in vitro.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
While awake and in the supine position, each subject turned his or her head 30–35 degrees to the left as measured by protractor. Each study subject was placed in moderate (10–15 degrees) Trendelenburg position to increase the size of the internal jugular vein (18). No respiratory maneuvers were used to increase the size of the internal jugular vein. If the landmark was not easily discernable, it was accentuated by having subjects try to lift their heads against resistance created by the placement of the hand of an investigator on their foreheads. The central landmark, defined as the apex of the triangle formed by the two heads of the sternocleidomastoid muscle as the sides of a triangle and the clavicle as the base, (12) was highlighted with an ink marker (Fig. 1; panel a). Subjects in whom the landmark could not be defined were not studied.
|
Two-dimensional ultrasound images of the neck anatomy were observed using a 7-MHz, 7V3c probe (10 mm footprint) and a Siemens-Acuson Sequoia® echocardiography machine. Throughout the study the same investigator (PB) placed the center of the probe on a subject's skin at the apex of the central landmark and the probe was aimed towards the subject's ipsilateral nipple, at approximately a 30 degree angle to the coronal plane and in a manner simulating how a syringe and needle would be placed for central venous catheterization (Fig. 1; panels b, c). The investigator placing the probe was blinded as to the image being generated. Another investigator operated the ultrasound machine. If necessary, the unblinded investigator would ask the investigator holding the probe to apply more gel to improve image quality. This was the only type of feedback given to the person placing the probe. Once image quality was adequate the vertical cursor, which served to delineate the path a needle would take as described by Lieberman et al. (19) was placed in the image. The image was subsequently recorded (Fig. 1; panel d). Ultrasound images, an example of which is in Figure 2, were collected for later analysis.
|
We also validated the simulation method used in vitro. We used an ultrasound simulation phantom (Blue PhantomTM; Advanced Medical Technologies, LLC), which is a gel block with several blood vessel-like structures internally. We held the ultrasound probe as described above (with the probe held at a slight angle as if aiming it towards the ipsilateral nipple in vivo) and obtained ultrasound images of one of the blood vessels in the phantom and then placed the vertical digital cursor in the image. We then placed an Echotip® percutaneous entry needle (Cook Inc., Bloomington, IN) into the phantom in line with the probe handle to determine if the ultrasound simulated needle path was the same as an actual needle path. We repeated this exercise 6 times with the probe handle at similar but not identical angles to the skin.
We sought to determine the accuracy of the central landmark. Accuracy is comprised of two elements: precision and bias. A procedure is precise if on repeating, it produces the same result with little scatter. A procedure is not biased if it does not consistently deviate from the true value in anything other than a random manner. Accuracy of the central landmark was defined in every subject depending on whether the projected needle path, as depicted by the path of the cursor, intersected the middle 80% of the IJ vein and did not intersect the carotid artery (Fig. 1; panel d). The length of a perpendicular line drawn from the center of the vein to the cursor (simulated path of a needle) was determined in every subject. This distance, multiplied by 2, when <0.8 x diameter of the vein, indicated if the cursor was within the middle 80% of the diameter. IJ vein diameter was also measured.
Results were depicted by constructing a plot that compared the distance between the projected needle path and the actual center of the IJ vein by drawing a perpendicular line from the center of the IJ vein to the simulated needle path for all study subjects. The distance was considered "positive" if the simulated needle path was medial to the center of the vein and was "negative" if the simulated needle path was lateral to the center of the vein. The scatter in this distance represents the precision. The bias of the landmark approach was obtained by calculating the mean of the distances measured as described above and in a manner similar to Bland and Altman (20). The significance of the bias was determined using a two-sided sign test.
Univariate logistic regression analysis was performed to assess the impact of obesity on the accuracy of the landmark approach. Obesity, defined as a body mass index more than or equal to 30 (21), was treated as a categorical variable. Two separate analyses were performed. In the first analysis, the dependent variable was whether the simulated needle path failed to intersect the middle 80% of the vein. In the second analysis, the dependent variable was whether or not the simulated needle path intersected the carotid artery. Statistical analyses were performed in STATA SE 8.2 (STATA Corporation, College Station, TX). All statistical tests were two-sided. Results are reported as mean ± sd unless otherwise specified.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
|
Data were evaluable in 104 of 107 subjects (54 patients and 50 volunteers). Fifty percent of the study subjects were female; 23 of 54 patients and 29 of 50 volunteers were female. The whole study group was 47.7 ± 21 (range, 18–93) yr old. Patients were 63 ± 15 (range, 21-93) and volunteers were 30.7 ± 11 (range, 18–67) yr old. Twenty-five percent (19 of 53 patients and 7 of 50 volunteers) of the study subjects were obese. Body surface area was 1.94 ± 0.26 (range, 1.38–2.73) and 1.82 ± 0.22 (range, 1.36–2.4) M2 for patients and volunteers, respectively. In one subject the central landmark could not be discerned and in two subjects data sets were incomplete. There were no complications or adverse events in any of the subjects.
RIJ vein diameter ranged from 2.5 to 61 mm (median, 17.3; interquartile range, 14.2–20.6). The central landmark did not identify the middle 80% of the lumen of the IJ vein in 34% of subjects (95% confidence interval [CI], 25% to 44%). Similarly, a needle directed using the central landmark would have taken a path in line with the right carotid artery in 26% of subjects (95% CI, 18% to 35%). Both events occurred in 20% of subjects (95% CI, 13% to 29%). The central landmark had a medial bias of 3.7 mm (95% CI, 2.7 to 4.8). Precision, as suggested by Bland and Altman (20) to be ± 2 standard deviations, was 10.5 mm. The central landmark resulted in a simulated needle path that was more often (77 of 104 subjects) medial to the center of internal jugular (P < 0.001; Fig. 4).
|
Compared with non-obese patients the simulated needle path based on the landmark technique was more likely to miss the IJ vein (odds ratio, 3.11; P < 0.016) and more likely to intersect the carotid (odds ratio, 3.03; P < 0.024) in obese patients.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
We found the central landmark to be significantly imprecise. The combination of imprecision and bias helps to explain difficulties that may be encountered during central venous catheterization with this approach. Variations in the size of the RIJ vein and the location of the carotid artery (23) will compound this problem. Whether central venous catheterization is performed using anatomic landmarks or ultrasound, the anatomical variations in the size and position of both the IJ vein and carotid artery should be appreciated.
We also found that the central landmark technique is more inaccurate in obese patients. In light of the prevalence of obesity in the United States, this finding may encourage clinicians to use ultrasound guidance for central venous catheterization in obese patients. The specific factors contributing to the greater inaccuracy of the central landmark in obese individuals may merit further study.
Many clinicians continue to perform central venous catheterization utilizing surface landmarks. Although English et al. (24) did report a small (<1%) carotid puncture rate, others have found more frequent rates (16). Denys and Uretsky (17) concluded that central venous catheterization should be anticipated to be difficult in a significant number of patients because in their study population the RIJ vein was not visualized (presumably thrombosed) in 2.5% of patients; the RIJ vein was unusually small in 3%; the RIJ vein was overlying the carotid artery in 2%, and the RIJ vein was described as "outside of the predicted path of the landmark" in 5.5%. Gordon et al. (14) found the RIJ vein to be medial to the carotid artery in 5.5% of patients.
Denys and Uretsky (17) reported a less frequent incidence than we found of both "missed" IJ veins and "hit" carotid arteries. Differences in methodologies may explain the discrepancies. In the Denys and Uretsky (17) report the "expected location of the internal jugular vein" was a subjective estimate that was neither blinded nor specifically quantified. Thus, investigator bias cannot be eliminated. Importantly many, but not all subjects in Denys and Uretsky's study performed a Valsalva maneuver (many subjects were awake), which significantly enlarged the IJ vein, while others were placed in Trendelenburg positioni. Finally, the degree of head turning in the study by Denys and Uretsky was not completely controlled.
Troianos et al. (22) obtained ultrasound images in 1136 patients and found the RIJ vein to overlie most (75%) of the carotid in 54% of patients and that this finding was more common in older (>60 yr) patients. The authors concluded that carotid puncture was a common risk of central venous catheterization if the cannulating needle traversed the RIJ vein. Because the IJ vein frequently overlies the carotid artery, with the carotid artery usually being behind the medial aspect of IJ vein, the bias of the central landmark approach will increase the risk of carotid puncture.
Other landmark-based techniques have been proposed (8,9) but also are unlikely to allow clinicians to address common causes of catheterization failure (e.g., small vein) or eliminate certain complications (e.g., carotid puncture). Techniques based on respiratory jugular venodilatation also have similar limitations (6,25). It is common for practitioners to palpate the carotid artery during central venous catheterization. Metz et al. (26) reported that techniques incorporating carotid palpation resulted in a reduced, but still significant, simulated incidence of carotid puncture (13% versus 27%). Oda et al. (7) in an unblinded uncontrolled study described a para-carotid approach and reported a success rate of 96% with 5 carotid punctures in 456 patients. However, some reports mention anecdotally that palpation tends to decrease the lumen size of the IJ vein. Bazaral and Harlan (27) mention that "palpation produces anatomical changes that make cannulation difficult." Palpation can not identify variations in anatomy, such as when the carotid artery is posterior to the IJ vein. The influence of palpation on central venous catheterization, both with and without the use of ultrasound merits further study.
The shape of the central landmark was very variable across study subjects and this likely contributed to its inaccuracy. We were able to feel and trace the central landmark in all but one subject. The few times that defining the landmark was difficult, having the patient maintain his head in the turned position while attempting to lift his head against a hand held on the forehead, helped to accentuate the landmark. We believe that delineating the landmark before the induction of anesthesia and eliminating muscle tone contributed to a successful appreciation of the landmark.
This study has potential limitations. First, we simulated needle path. We did not consider it ethical for us to achieve our study goals by actually passing a needle instead of simulating the needle path because most catheterizations in our cardiac surgical patients have been assisted with the use of ultrasound for several years. Validity of our model was based on the following: 1. Others have used the same simulation technique when studying certain aspects of central venous catheterization (19,26); 2. The "cursor" on ultrasound images is used to guide procedures such as pericardiocentesis (28); 3. We validated our methods in vitro. Nevertheless, the potential for inaccuracies in the 1–2 mm range to influence our results cannot be completely eliminated.
Another potential design weakness is that most of our study subjects were at first the adult volunteers and that patients were mostly studied afterward. This occurred because numerous volunteer subjects could be studied in one day but only one or two patients could be studied on any given day. Thus, enrollment for the volunteers proceeded at a more rapid pace than for patients. We included nonsurgical volunteer subjects in our study population so as to be able to determine if there was any influence of age on our findings, which there was not.
The risk of carotid puncture we found is more frequent than that published in the literature. In adults, this risk is often reported to be in the 5%–10% range but more frequent carotid puncture rates, especially in children or obese individuals, have been reported. Metz et al. (26) also found a risk of carotid puncture similar to ours (13%–27%). Some (approximately 6%) of our simulated carotid punctures may not occur clinically because they would occur only after a needle traversed the IJ vein lumen. This fact contributes to our increased simulated incidence of carotid puncture but highlights how carotid puncture remains a risk when it lies posterior to the IJ vein. Also, some needles aimed at the carotid artery may "roll off" the vessel and not actually puncture it.
Variations in technique, such as the degree of Trendelenburg position or head turning used or whether breath holding was used could influence results. However, turning the head to more than 30–40 degrees has been demonstrated to be disadvantageous (19,29). Although cyclical variations in intrathoracic pressure with spontaneous or positive pressure ventilation do influence central venous pressures and IJ vein cross-sectional area, they probably influence IJ vein diameter to a lesser degree. However, we did not evaluate this. Our findings should be applied with caution to the left IJ vein.
In summary, bias of the central landmark for RIJ vein catheterization leads clinicians to insert needles more medially than desired. Inaccuracy of the central landmark approach may, on first needle pass, lead to failure to identify the lumen of IJ vein in up to a third of patients and/or the risk of carotid puncture in up to a quarter of patients. These central landmark limitations are increased in obese patients. If the central landmark approach is used the bias of the landmark should be considered.
|
Footnotes |
|---|
Presented, in part, at the Annual Meeting of the Society of Cardiovascular Anesthesiologists, Baltimore, Maryland, May 16th, 2005.
All funding for the project came from within the Department of Anesthesiology, University of Rochester.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  P. L. Bailey, L. G. Glance, M. P. Eaton, B. Parshall, and S. McIntosh A Survey of the Use of Ultrasound During Central Venous Catheterization Anesth. Analg., March 1, 2007; 104(3): 491 - 497. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Ye, T. Zhu, and J. Liu Central Landmark for Central Venous Catheterization Anesth. Analg., January 1, 2007; 104(1): 216 - 216. [Full Text] [PDF]
|
|
|
|
|
|
P. Bailey, E. Whitaker, and L. Glance Central Landmark for Central Venous Catheterization Anesth. Analg., January 1, 2007; 104(1): 216 - 217. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
