THE internal jugular vein (IJV) is a route commonly used to access the central circulation because of its accessibility during surgery and predictable anatomic location. Although cannulation of the IJV, using anatomic landmarks, is associated with a 95% success rate, 
researchers in recent studies demonstrated improved success and fewer complications when ultrasound guidance is used to facilitate cannulation. [2-4]
Anatomic variation or unreliability of the external landmarks was implicated as the cause for difficulty in cannulation without ultrasound. 
An anatomic relation in which the IJV overlies the carotid artery (CA) has accounted for unintentional CA puncture. [2,4]
The purpose in this study was to identify the anatomic relation of the right IJV and CA in the direction of a cannulating needle using ultrasound imaging in 1,136 patients and to determine which patient characteristics are associated with an overlying (instead of laterally positioned) IJV.
Materials and Methods
With approval of the Research and Human Rights Committee of The Mercy Hospital of Pittsburgh, and written informed consent, 1,136 patients admitted for surgery were studied prospectively. Patients with prior neck surgery were excluded from the study.
The ultrasound device (Site Rite, Dymax, Pittsburgh, PA) consisted of a 7.5-MHz transducer, a two-dimensional image display, and a Polaroid(R) camera (Polaroid Corp., Cambridge, MA), modified for close-up photography. Awake patients were placed supine on a stretcher parallel to the floor without a pillow, and their heads were rotated as far to the left as was comfortable. Imaging of the right IJV was performed at the apex of the angle formed by the division of the sternocleidomastoid muscle. The needle guide attached to the transducer was positioned in the direction of a cannulating needle entering the skin at the apex of the angle and directed toward the ipsilateral nipple (Figure 1
). The imaging plane of the probe was directed in the same plane as a cannulating needle, and not in a coronal plane. A 7.6 x 7.6 cm photograph was taken of the ultrasound image of the CA and IJV.
Patient characteristics, including age, height, weight, body surface area, and history of chronic obstructive pulmonary disease, myocardial infarction, or congestive heart failure, were recorded.
All photographs were scored later independently by three investigators who were blinded to patient characteristics and the scoring by the other investigators. Scoring was defined as follows:
0 = the IJV was positioned completely lateral to the CA on the image display;
1 = the IJV overlapped up to and including 25% of the diameter of the CA;
2 = the IJV overlapped more than 25% and up to and including 50% of the diameter of the CA;
3 = the IJV overlapped more than 50% and up to and including 75% of the diameter of the CA; and
4 = the IJV overlapped more than 75% of the diameter of the CA.
This scoring system was used to identify patient anatomy that allows for carotid puncture by a needle that traversed the IJV before entering the CA. An increased likelihood of carotid puncture in this manner would be expected with a greater score.
Patients were eliminated from statistical analysis if the image quality was poor or if scoring was disparate among the three scorers (2 or greater difference in score). If the three independent scores differed by one, the score of the two concurring investigators was used for statistical analysis.
Data were analyzed by chi-square analysis to compare incidence of score between different age groups. Both one-way analysis of variance and Kruskal-Wallis one-way analysis of variance by ranks were used to examine whether age, weight, height, vein size, artery size, and history of chronic obstructive pulmonary disease, myocardial infarction, or congestive heart failure were different by score category. Linear regression was used to test the relation between vein size and other continuous variables. A P value less than 0.05 was considered to be significant.
The Polaroid(R) photographs of 1,136 patients were reviewed by three investigators (CAT, RJK, JRP). Poor image quality and disagreement in score by 2 or greater eliminated 127 patients from statistical analysis. Among the remaining 1,009 patients, 54% of these patients received a score of 4 (more than 75% of the CA is overlaid by the IJV; Table 1
). The only demographic data category predictive of a score of 4 was age (P < 0.05). Patients with more advanced age were more likely to have a score of 4. Sixty-four percent (64%) of patients aged 60 yr or older had a score of 4, whereas only 47% of patients younger than 60 yr had a score of 4 (P < 0.05). A greater percentage of the patients younger than 40 yr had scores less than 3 (46% of these younger patients compared with only 24% of patients aged 60 yr or older; P < 0.05).
Vein size (both width and antero-posterior diameter) was both positively related to score (P < 0.01). Patients with larger veins had a higher score. There was no correlation between weight, height, or age with vein size. There was no difference in vein size with respect to the patient's gender or a history of myocardial infarction, congestive heart failure, or chronic obstructive pulmonary disease (Table 2
Cannulation of the IJV is commonly performed using visual and palpable anatomic landmarks. [6,7]
A variety of techniques have been used to access the IJV, without demonstration of superiority of one technique over another. 
Anatomically, the CA is posterior and medial to the IJV, and techniques that incorporate palpation of the CA pulse result in a lower incidence of CA puncture. 
Carotid artery puncture can occur either primarily (needle placement directly into CA lumen), or secondarily (after the cannulating needle transverses the IJV 
). The ease with which the IJV is compressed accounts for the initial undetected entry into the IJV. 
The anatomic position of the IJV is described classically as lateral to the CA. 
This anatomic relation describes the relation of these structures in the coronal plane, not in the directional plane of the cannulating needle with the head turned to the contralateral side (Figure 2
). This may explain why our study revealed a greater proportion of patients than expected whose IJV was positioned anterior rather than lateral to the CA. If the IJV is anterior to the CA, undetected entry (and exit) through a compressed IJV may result in puncture of the CA, as reported previously. [2,4]
In ultrasound-simulated cannulation techniques, Metz et al. 
noted the consistency of the anatomic relation between the IJV and CA (the IJV being anterior and lateral to the CA). They suggested that because the IJV was found consistently to be lateral to the CA, the CA pulse can be used as a primary or alternative landmark. Their study population, however, consisted of hospital employees, aged 24-38 yr. This is consistent with the current study, which demonstrated a more laterally positioned IJV in younger patients. The current study examined the anatomic relation in a variety of age groups and demonstrated that although the IJV position is more often overlying rather than lateral to the CA (54% of all patients), a greater proportion of patients older than 60 yr have this anatomic relation (Table 1
). If this anatomic relation is present, initial attempts of IJV cannulation would be unsuccessful if the CA was the sole landmark. Subsequent repositioning of the cannulating needle medially could result in CA puncture after undetected entry into a compressed IJV. The ease of vein compression with minimal external pressure or needle advancement is apparent when ultrasound is used to facilitate cannulation of the IJV. [2,4,9]
In a recent study of 15 hospital workers, aged 18-60 yr, the effect of head position on the relative positions of the CA and IJV was examined. Sulek and colleagues 
showed greater overlap between the vessels when the head was rotated 80 degrees compared with rotation of 0 degree and 40 degrees. This is consistent with our study, in which we demonstrated overlap when patients were placed in a cannulating position with their head rotated to the opposite side. However, why did our study reveal more overlap between the two vessels in the older patients? One would expect that older patients would not be able to rotate their head to the same degree as younger patients. If head rotation alone determines the degree of overlap, less overlap would be expected in older patients. The Sulek et al. study, with its sample size of only 15, did not examine the effect of age on the degree of overlap. Unfortunately, we did not measure the degree of head rotation during our study to comment on whether head rotation affects the amount of overlap between the IJV and CA. In another study, where ultrasound-simulated cannulation was used, Metz et al. 
found that head rotation did not affect the likelihood of cannulating the IJV. Their subjects, however, were all younger than 40 yr. Our observation of greater vessel overlap in older patients is likely due to the fact that the common CA becomes elongated and tortuous in older patients, presumably from arteriosclerosis. 
The effect of the "cannulating" position on the anatomic relation of the vessels is demonstrated by the magnetic resonance image in Figure 3
. The magnetic resonance image confirms what we observed with ultrasound. Namely, when patients are placed in a cannulating position with head turned to the contralateral side, there is often overlap between the IJV and CA.
If a score of 3 or 4 (> 50% of the CA is overlaid by the IJV) is considered a more "dangerous" anatomic relation for CA puncture, the results are impressive for all age groups (68% of all patients had scores of 3 or 4). Although patients younger than aged 60 yr had a statistically lower incidence of a score of 4, compared with older patients, the most common score in younger patients also was 4. Why is the CA puncture rate reported to be only 4% 
rather than 68%? There are two possible explanations. First, although 68% of our patients had the IJV overlying more than half the cross-sectional diameter of the CA, the amount of the IJV that did not overlie the CA was not examined. For example, a large IJV may overlie the majority of the CA, but, in addition, extend beyond the lateral edge of the CA. We did find that vein size was related positively to scores, because patients with larger veins tended to have greater scores. A needle directed toward the portion of vein that does not overlie the CA would have a lower risk of puncturing the CA than a needle directed toward the portion of vein that overlies the CA. Second, if a cannulating needle enters the IJV and is aimed at the CA, entry into the IJV may be detected before the needle exits the lumen through the posterior wall of the vein. Mangar et al. 
reported that access to the IJV is achieved 50% of the time with needle advancement and 50% during needle withdrawal. Carotid puncture is more likely to occur when the cannulating needle punctures the posterior wall of the IJV before blood is aspirated. Use of a "finder" needle, which is smaller and sharper, may not compress the IJV to the same degree as a larger cannulating needle, minimizing the likelihood of penetrating the posterior wall of the IJV. Many practitioners mistakenly believe that the practice of using a "finder" needle will protect them from puncturing the CA in lieu of using ultrasound guidance. Even if the cannulating needle follows the same track as the finder needle, IJV compression may occur with the cannulating needle because of its large diameter.
Another issue is that because a significant degree of vessel overlap occurs at the level of the apex of the sternocleidomastoid muscle division, is there another level of needle entry that reliably provides a more favorable orientation of the IJV and CA? Our study design did not include ultrasound examination at various levels because we were most interested in describing the anatomy at the level most commonly used. Examination at different levels may have provided more insight into a more optimal approach to the IJV using surface landmarks. Our purpose in the study, however, was not to find the optimal entry point, but rather to redirect conventional thinking and teaching that describes the anatomic relation of the IJV as lateral to the carotid. We believe that the optimal approach to the IJV should be by ultrasound guidance. When an IJV overlying the CA is identified, the ultrasound transducer is moved to a level that provides a more favorable anatomic relation.
The superiority of ultrasound-guided cannulation compared with landmark-guided cannulation is clear. Overall success rate, first pass cannulation, number of needle advances, and carotid puncture rate all were shown to favor the ultrasound-guided technique in several randomized, controlled studies. [2-4,14]
The valuable role of ultrasound in cannulation of the IJV is apparent, particularly for patients whose external anatomic landmarks are unclear or distorted, [5,15]
for patients for whom cannulation could not be accomplished readily using external landmarks, [2,3,5]
and where avoidance of CA puncture is especially important (previous carotid surgery or coagulopathy). 
Our images were obtained from patients with normal surface anatomy. Based on appearance and palpation alone, it is impossible to predict which patients will have a "safe" anatomic relation and which will have a "dangerous" anatomic relation (IJV overlies CA). Our observations explain in part why studies comparing ultrasound with landmark-guided cannulation all favor the ultrasound-guided technique. [2-4,14]
The authors thank Dr. David Jobes and Dr. Norig Ellison, for editorial assistance; Dr. Roberta Pasqual, for protocol execution; Dr. Frank D'Amico, for statistical analysis; Dr. M. Reza Raji, for radiologic expertise; John Schulz, for illustration; Peggy Flynn, for secretarial assistance; and Thomas Mathie, for help with references.
1. Jobes DR, Schwartz AJ, Greenhow DE, Stephenson LW, Ellison N: Safer jugular vein cannulation: Recognition of arterial puncture and preferential use of the external jugular route. ANESTHESIOLOGY 1983; 59:353-5.
2. Troianos CA, Jobes DR, Ellison N: Ultrasound-guided cannulation of the internal jugular vein: A prospective, randomized study. Anesth Analg 1991; 72:823-6.
3. Mallory DL, McGee WT, Shawker TH, Brenner M, Bailey KR, Evans RG, Parker MM, Farmer JC, Parillo JE: Ultrasound guidance improves the success rate of internal jugular vein cannulation. Chest 1990; 98:157-60.
4. Denys BG, Uretsky BF, Reddy PS: Ultrasound-assisted cannulation of the internal jugular vein: A prospective comparison to the external landmark-guided technique. Circulation 1993; 87:1557-62.
5. Denys BG, Uretsky BF: Anatomic variations of internal jugular vein location: Impact on central venous access. Crit Care Med 1991; 19:1516-9.
6. English ICW, Frew RM, Pigott JF, Zaki M: Percutaneous catheterisation of the internal jugular vein. Anaesthesia 1969; 24:521-31.
7. Defalque RJ: Percutaneous catheterization of the internal jugular vein. Anesth Analg 1974; 53:116-21.
8. Metz S, Horrow JC, Balcor I: A controlled comparison of techniques for locating the internal jugular vein using ultrasonography. Anesth Analg 1984; 63:673-9.
9. Ellison N, Jobes DR, Troianos CA: Internal jugular vein cannulation. Anesth Analg 1994; 78:198.
10. Goss CM: The veins, Gray's Anatomy. Philadelphia, Lea & Febiger, 1973, pp 697-8.
11. Sulek CA, Gravenstein N, Blackshear RH, Weiss L: Head rotation during internal jugular vein cannulation and the risk of carotid artery puncture. Anesth Analg 1996; 82:125-8.
12. Teal JS, Rumbaugh CL, Bergeron RT, Segall HD: Lateral position of the external carotid artery: A rare anomaly? Radiology 1973; 108:77-81.
13. Mangar D, Turnage WS, Mohamed SA: Is the internal jugular vein cannulated during insertion or withdrawal of the needle during central venous cannulation? Anesth Analg 1993; 76:1375.
14. Alderson PJ, Burrows FA, Stemp LI, Holtby HM: Use of ultrasound to evaluate internal jugular vein anatomy and to facilitate central venous cannulation in paediatric patients. Br J Anaesth 1993; 70:145-8.
15. Troianos CA, Savino JS: Internal jugular vein cannulation guided by echocardiography. ANESTHESIOLOGY 1991; 74:787-9.
© 1996 American Society of Anesthesiologists, Inc.