Ultrasound imaging of the axillary vein—anatomical basis for central venous access

S. Galloway and A. Bodenham

Department of Anaesthesia, Leeds General Infirmary, Great George Street, Leeds LS1 3EX, UK

Corresponding author. E-mail: andy.bodenham@leedsth.nhs.uk

Accepted for publication: March 19, 2002


    Abstract
 Top
 Abstract
 Introduction
 Surface landmark techniques
 Applied anatomy
 Methods
 Results
 Discussion
 Conclusion
 References
 
Background. The central veins that are usually cannulated are the jugular, subclavian, femoral and brachial. If subclavian catheterization is difficult using surface landmark techniques, we now use ultrasound to catheterize the infraclavicular axillary vein. This approach is not widely used and the ultrasound appearance has not been formally described. We examined the anatomical relationships of the axillary vessels to guide safe cannulation of the axillary vein.

Method. In 50 subjects, we used ultrasound to examine the infraclavicular regions from below the mid-clavicular point and at 2 cm and 4 cm further laterally (described as the middle and lateral points) with the arms at 0°, 45° and at 90° abduction. We took measurements at each point, with the artery and vein seen in cross-section. The depth from the skin, vessel diameters and the distance between the vessels was measured. The amount of overlap was scaled from 0 (no overlap) to 3 (complete overlap). We also recorded (if visible) the distance between the rib cage and axillary vein. A longitudinal image of the vein was also obtained. Angle of ascent (in relation to the skin), length and depth of the vein was measured.

Results. Axillary vessels were seen in 93% of images. The mean depth from skin to vein increased from 1.9 cm (range 0.7–3.7 cm) medially to 3.1 cm (1.1–5.6 cm) laterally. The venous diameter decreased from 1.2 cm (0.3–2.1 cm) medially to 0.9 cm (0.4–1.6 cm) laterally. The arterio–venous distance increased from 0.3 cm to 0.8 cm. Median arterio–venous overlap decreased from 2/3 (mode 3/3) to 0 (0). The distance from rib cage to vein increased from 1.0 cm to 2.0 cm.

Conclusion. The axillary vein is an alternative for central venous cannulation and we present an anatomical rationale for its safe use. Less arterio–venous overlap and a greater distance between artery and vein and from vein to rib cage should provide an increased margin of safety for central venous cannulation.

Br J Anaesth 2003; 90: 589–95

Keywords: equipment, catheters, venous; measurement techniques, ultrasound


    Introduction
 Top
 Abstract
 Introduction
 Surface landmark techniques
 Applied anatomy
 Methods
 Results
 Discussion
 Conclusion
 References
 
Central venous catheterization is an important aspect of patient management in many clinical circumstances in anaesthesia, longer term venous access and intensive care. The jugular, subclavian, femoral and brachial veins are most frequently used. We often catheterize the subclavian vein for insertion of Hickman catheters and intensive care. If surface landmark techniques are not helpful, we now use ultrasound to cannulate the infraclavicular axillary vein. Informal observations suggested that this approach had anatomical advantages. The use of 2D ultrasound for the axillary approach is less well recognized and has not been formally described.

We set out to examine the anatomical relationships of the axillary vessels and surrounding structures to guide the safe cannulation of the axillary vein. We wished to confirm that a more lateral approach would have anatomical advantages giving a greater safety because the vein lies at a greater distance from the artery and rib cage, or because the overlap between the vessel changes on moving from medial to lateral. This was purely an observational study.


    Surface landmark techniques
 Top
 Abstract
 Introduction
 Surface landmark techniques
 Applied anatomy
 Methods
 Results
 Discussion
 Conclusion
 References
 
Most infraclavicular approaches to central venous catheterization are directed at the subclavian vein. However, many clinicians who gain confidence with subclavian vein cannulation move more laterally and puncture the skin at the midclavicular point. This probably represents skin puncture over the axillary vein but either the subclavian or axillary vein may be punctured depending on the distance between the skin and vein puncture sites.

A blind landmark-based technique for cannulating the axillary vein was suggested in 1987 by Nickalls and colleagues.1 They studied cadavers and suggested a complex landmark-based approach to the infraclavicular portion of the axillary vein. They described an insertion point three fingerbreadths below the coracoid process slightly lateral to the lateral border of pectoralis minor. The needle is aimed towards the point below the medial end of the clavicle where the space between the clavicle and the thorax just becomes palpable. This is approximately the junction of the medial quarter and lateral three-quarters of the clavicle. Despite this apparently complicated technique they claimed success in 13 out of 14 patients. In a larger study Taylor and Yellowlees2 used a slight modification to this technique using a landmark-based method. The success rate for cannulation of the axillary vein was 96% (similar to that reported for subclavian vein cannulation by Mogil and colleagues3 in 1967).

Neither of the blind surface landmark techniques seems to be widespread. This may be because anatomical variability may reduce reliability. The anatomy of the veins of the upper limbs often varies.4 Cadaver-based studies show that arm position affects the position of the axillary vein1 5 and venogram-based studies show sex differences in axillary vein straightness and diameter.6 The axillary vein and artery often overlap.7 This variation in relationship between the axillary artery and axillary vein is clinically important because axillary vein puncture may be associated with simultaneous arterial puncture if a landmark-based technique is used, with complications.

We used ultrasound to show anatomical variation in a study of femoral vascular anatomy.8


    Applied anatomy
 Top
 Abstract
 Introduction
 Surface landmark techniques
 Applied anatomy
 Methods
 Results
 Discussion
 Conclusion
 References
 
The anatomy of the axillary and subclavian veins has been described in detail in many standard anatomy texts (Fig. 1). The axillary vein is the continuation of the basilic vein and extends from the outer border of teres major to the outer border of the first rib. The curvature of the first rib means that anteriorly the axillary–subclavian junction is very medial. Its route from deep in the axilla takes it on a course that runs from lateral to medial, inferior to superior and posterior to anterior. The axillary vein is crossed immediately anteriorly by pectoralis minor, which divides the vein topographically into three parts, namely proximal, posterior and distal to pectoralis minor. The structures that lie posterior to the axillary vein are of obvious concern if this vein is to be used for central venous access. In its more medial portions the axillary vein (like the subclavian vein) is bordered posteriorly by the anterior rib cage. However in the vein’s middle portion the rib cage ‘falls away’ inferiorly to leave a greater gap between the vein and the rib cage. In addition the steeper angle of descent of the rib cage increases the chance of a needle sliding off rather than puncturing it. Further laterally, the rib cage is no longer posterior to the axillary vein. In fact, here there are no particularly vulnerable structures posterior to the vein. Examination of cadavers shows that a misplaced needle passing through the axillary vein will travel posteriorly through the axillary fat, through the serratus anterior muscle, through the subscapularis muscle and on to the scapula. The vein runs on the medial side of the axillary artery to this point where it continues as the subclavian vein. The only potential problem with the very lateral sections of the axillary vein is the proximity of the brachial plexus. The vein and plexus travel in a neurovascular bundle along with the axillary artery, which places the plexus close to the target of venous cannulation. Most often, however, the brachial plexus is placed posteriorly with regard to the axillary vein and is thus not often in the way. In fact, this anatomical relationship is not significantly different from that of the subclavian vein and brachial plexus.



View larger version (108K):
[in this window]
[in a new window]
 
Fig 1 Simplified diagram of the anatomy of the infraclavicular region, redrawn from many sources. The pectoralis minor muscle has been reflected to facilitate the view. Note the area lateral to the rib cage, which would contain only axillary fat. Vein=axillary vein; Artery=axillary artery; PM=pectoralis minor muscle; Fat=axillary fat.

 

    Methods
 Top
 Abstract
 Introduction
 Surface landmark techniques
 Applied anatomy
 Methods
 Results
 Discussion
 Conclusion
 References
 
Approval for the study was obtained from the local research ethics committee. Fifty patients admitted to the general intensive care unit (ICU) or general surgical wards were studied. Patients were only excluded from the study if they had a subclavian vein cannula in situ or if they had recently suffered clavicular trauma. Consent was obtained from the patients before their examination or assent from their relatives if the patient was unable to give consent. Patients were not given i.v. fluids before the study. Simple data were collected for each patient, including age, height, weight and primary diagnosis. Height was measured but weight was estimated because of the difficulty in measurement in sicker patients. Examinations were carried out using a portable ultrasound machine (Scanner 100, Pie Medical Ltd, Crawley, UK) with 5/7.5 MHz dual-frequency curvilinear and 7.5 MHz linear probes. Patients were examined in the supine position. The right and left infraclavicular regions were examined with the arms in the neutral position, at 45° and at 90° abduction. Examinations were made just below the mid-clavicular point (0 cm) and at 2 cm and 4 cm lateral to that point following a line that kept the vessels in the centre of the image (Figs 2 and 3). Shoulder position (elevation or depression) was not controlled. All examinations were made using the least amount of pressure required between probe and skin in order to obtain an optimal image. Measurements were taken at each point, with the vessels seen in cross-section. The depth of the axillary artery and axillary vein below the skin, the vessel diameters, the distance between the vessels and the extent to which the vessels overlapped were measured with software present in the ultrasound machine. The overlap of one vessel over the other was estimated on an arbitrary scale with 0/3 being no overlap, 3/3 being complete overlap, and 1/3 and 2/3 being one-third and two-thirds of the artery overlapped from above by the vein, respectively. This was measured from directly anteriorly. We also recorded whether the rib cage was seen in the same image as the vein. If it was seen then the distance between the rib cage and axillary vein was recorded.



View larger version (126K):
[in this window]
[in a new window]
 
Fig 2 A normal left axillary venogram. Contrast has been injected into a distal vein. Note the irregular outline of the vein caused by valves. Some contrast is also seen in the cephalic vein. The axillary vein is seen draining into the brachiocephalic system. Note also the incidental right-sided central venous catheter. The line a–a is the midclavicular point where the scan for Figure 3a was taken. The line b–b is 2 cm lateral to the midclavicular point and is where the scan for Figure 3b was taken. The line c–c is 4 cm lateral to the midclavicular point and is where the scan for Figure 3c was taken.

 


View larger version (62K):
[in this window]
[in a new window]
 
Fig 3 Transverse ultrasound images of the axillary vessels taken with a dual-frequency 5/7.5 MHz curvilinear ultrasound probe. A is a scan from immediately below the midclavicular point (0 cm); B is a scan from 2 cm lateral; C is a scan from 4 cm further lateral. Note that from A to B to C the size of the vein decreases, the distance between the vessels increases, the arterio–venous overlap decreases and the distance between the vein and rib cage increases. Vein=axillary vein; Artery=axillary artery; RC=rib cage.

 
Vessel diameters and the degree of overlap were recorded using the curved probe, with the patient in a 10° head-down tilt. Subsequently, a longitudinal image of the axillary vein was recorded with the linear probe, and the length of the section of vein, maximal depth and angle of ascent of the vein in relation to the skin surface were recorded (Fig. 4). Simple descriptive statistics were used to describe the anatomy.



View larger version (160K):
[in this window]
[in a new window]
 
Fig 4 Longitudinal ultrasound image of the axillary vein taken with a 7.5 MHz linear probe. Note the angle at which the vein rises up from the axilla, compared with the skin surface.

 

    Results
 Top
 Abstract
 Introduction
 Surface landmark techniques
 Applied anatomy
 Methods
 Results
 Discussion
 Conclusion
 References
 
The mean age of the patients studied was 64.3 (range 30–90) yr. The sample consisted of 15 females and 35 males. Mean weight and height were 73 (SD 13) kg and 169 (10) cm, respectively. The mean clavicular length was 15 (range 12–19) cm.

Of the 50 patients scanned, 41 scans (82%) were complete in terms of visibility of the vessels. Sometimes both vessels could not be seen adequately. The mean BMI of the patients with incomplete scans was 27.2, compared with 25.2 for the patients in whom scanning was complete. In those patients with incomplete scans, it was usually the more lateral scans that were difficult or inconclusive.

The relationships between the various tributaries to the subclavian vein varied. Most often (but not always) the cephalic vein was smaller than the axillary vein and joined it laterally. However, in a small number of cases the cephalic vein arose more medially, appearing in the most medial or middle slices. This occurred in approximately 12% of subjects scanned (Fig. 5).



View larger version (144K):
[in this window]
[in a new window]
 
Fig 5 Transverse ultrasound images of axillary vessels. Note the abnormal configuration of vessels. Ax=axillary vein; Art=axillary artery; Ceph=cephalic vein.

 
Abnormalities were seen in seven patients (14%). These were mainly unusual combinations of tributaries, but also included incompressible vessels in the absence of obvious thrombus. Thrombus was seen in six patients (12%) (Fig. 6). Only three of these patients had previously had central venous catheters inserted. The other three were idiopathic, perhaps related to long-term hospitalization or critical illness.



View larger version (118K):
[in this window]
[in a new window]
 
Fig 6 Transverse ultrasound images of axillary vessels. Note the thrombus visible in the axillary vein. This thrombus was incompressible when pressure was applied with the ultrasound probe. Throm=thrombus.

 
We looked for the brachial plexus but this could not be seen reliably. This is unsurprising with the probes that we were using.

The mean depths and diameters of the vessels for the different ultrasound slices and the mean arterio–venous distance are shown in Tables 1, 2, 3. The mean venous diameter with 10° head-down tilt was 1.3 cm. The mean arterial diameter with 10° head-down tilt was 0.8 cm.


View this table:
[in this window]
[in a new window]
 
Table 1 Mean depth from the skin, diameter of axillary vessels, arterio–venous distance, and rib cage to vein distance in cm, with arm in 0° abduction. Data are mean (range)
 

View this table:
[in this window]
[in a new window]
 
Table 2 Mean depth from the skin, diameter of axillary vessels, arterio–venous distance and rib cage to vein distance, in cm, with arm in 45° abduction. Data are mean (range)
 

View this table:
[in this window]
[in a new window]
 
Table 3 Mean depth from the skin, diameter of axillary vessels, arterio–venous distance and rib cage to vein distance, in cm, with arm in 90° abduction. Note that the rib cage was never visible in the lateral (4 cm) slice. Data are mean (range)
 
Degrees of overlap between the vein and artery are shown in Tables 4, 5, 6. The median arterio–venous overlap with the subject tilted 10° head down was 2/3 (range 0–3).


View this table:
[in this window]
[in a new window]
 
Table 4 Numbers of patients with vein overlapping artery at 0° abduction (percentages are given in parentheses)
 

View this table:
[in this window]
[in a new window]
 
Table 5 Numbers of patients with vein overlapping artery at 45° abduction (percentages are given in parentheses)
 

View this table:
[in this window]
[in a new window]
 
Table 6 Numbers of patients with vein overlapping artery at 90° abduction (percentages are given in parentheses)
 
The rib cage was seen in 388 of the 900 slices taken. Of these, 280 (72%) were seen in the 0 cm slice, 101 (26%) were seen in the 2 cm slice and 7 (2%) were seen in the 4 cm slice. Mean rib cage to vein distances are shown in Tables 13.

Longitudinal sections
The mean depth of the deepest part of the vein seen on the longitudinal scan was 2.7 cm. The mean angle between the skin and the vein was 13.3° (range 1–29°).


    Discussion
 Top
 Abstract
 Introduction
 Surface landmark techniques
 Applied anatomy
 Methods
 Results
 Discussion
 Conclusion
 References
 
Rationale
The subclavian route is useful for central venous access but it has a number of potential complications such as arterial puncture with possible haematoma or arterio–venous fistula, or pneumothorax.

Ultrasound cannot be used to show the subclavian vein as, by definition, this vessel is under the clavicle. It could be used for the supraclavicular approach although this is a less popular technique. Ultrasound aids accurate cannulation of many central veins, reducing complications.9 10 Use of ultrasound for axillary vein cannulation may also reduce complications, as suggested by Doppler studies.11

Safe effective cannulation of the axillary vein by the infraclavicular approach is attractive as it shares the advantages of the subclavian route (clean site, comfortable for patient). In addition, compared with the subclavian vein, there should be less risk of pneumothorax, haemothorax and chylothorax. In the event of inadvertent arterial puncture, a more lateral approach allows direct external pressure to be applied and easier surgical access (the same as for axillo-femoral bypass). Other patients who might benefit from a more lateral puncture include those with tracheostomies, recent sternotomy and those who have midline thoracic burns.

Ultrasound
We did not expect to see the brachial plexus. There is an inverse relationship between depth of penetration of ultrasound and image resolution. We compromised our image resolution in order to achieve adequate depth of penetration so that the vessels would be visible at the greater depths in the lateral sections. Study of these structures has been reported in a recent paper.12

We found considerable variation in the anatomy, which would not be appreciated with surface landmarks. This variation makes the use of ultrasound attractive when accessing these vessels. With other central veins, ultrasound can aid accurate cannulation and hence reduce complications.9 10

Practical application
We examined patients in the ICU as the study was time-consuming and therefore not appropriate for use in more urgent circumstances. In addition, the patients were ‘real’ rather than volunteers who could have had more favourable characteristics.

We could not see the vessels at all sites in all patients. In only one patient were no vessels visible in any of the scanned areas. Vessels were less easily imaged in larger patients. The mean BMI for the patients with incomplete scans was 27.2, compared with 25.2 for those with complete scans. Satisfactory scans were much harder to obtain from patients with oedema over the anterior thoracic wall. Surgical emphysema makes ultrasound of this region very difficult. It is also of interest that subclinical surgical emphysema was detected in some patients in this way. A permanent pacemaker made probe positioning difficult in one patient.

Overlap of the artery and vein
We found more vessel overlap in the more medial sections. This could cause problems. Needle advancement reduces venous diameter.13 With central venous cannulation, the first aspiration of blood often occurs on withdrawal of the needle,14 15 especially when a narrower gauge seeker needle is not used, probably because the vein is compressed and transfixed.16 This means that if the artery lies behind the vein, then it may be punctured, causing possible haematoma, haemothorax or even arterio–venous fistula. This is a good reason for a more lateral approach.

Depths and angles
We found that the vein would be deeper with a more lateral approach. The mean depth in the most lateral ultrasound slices was 3.2 (range 1.4–5.4) cm, 3.1 (1.3–5.1) cm and 3.1 (1.1–5.6) cm in the midclavicular, 2-cm and 4-cm slices, respectively. At the greater depth higher frequency ultrasound probes, which have limited depth range, will not provide a good image in obese or muscular subjects, and we found this to be a clinical problem.

This depth also requires a much steeper angle of approach with the needle when attempting to cannulate the vein. This increase in depth of the vessels and the more lateral position will affect the length of catheter required, up to 10 cm. This is not a problem for catheters which are cut to length, but for fixed-length catheters a 20 cm catheter (or longer) is likely to be required.

The longitudinal sections show the angle at which the vein rises up from the axilla. It is also worth noting that the skin at this point is also rising up from the axilla onto the anterior rib cage so that the angle of ascent is actually greater than that shown.

Rib cage
We found that the rib cage was less visible in the more lateral slices. In those patients whose rib cage is visible more laterally, the distance between the axillary vein and the rib cage is greater. In fact if we looked for the rib cage it was always visible but often not close to the vessels. This suggests an advantage over the subclavian vein, which is in much closer proximity throughout its length, hence the greater likelihood of pneumothorax with this traditional approach.

Subject position
Tilting the patient 10° head down had no effect on the arterial diameter. As expected, the diameter of the axillary vein was increased by head-down tilt. This increase from 1.2 cm to 1.3 cm is small but may be of clinical significance. More importantly, head-down tilt may increase the venous pressure, which would make cannulation easier and transfixion less likely. Tilting the patient 10° head down did not alter the overlap. The median value was 2/3 (2 for the untilted).

Arm position
There were no differences between the arm positions in terms of visibility of vessels. Generally, different degrees of abduction of the arm had only small effects. Vessel depths and diameters were largely unchanged with increased abduction. There was small but consistent decrease in the distance between the vein and the rib cage. It is not clear if this was caused by abduction, supination or both, as the movements were not distinguished. There does not appear to be any great benefit in the use of one arm position over another.


    Conclusion
 Top
 Abstract
 Introduction
 Surface landmark techniques
 Applied anatomy
 Methods
 Results
 Discussion
 Conclusion
 References
 
On moving laterally, the axillary vein and artery lie further apart and further away from the rib cage. This provides an anatomical reason for the use of the axillary vein for central venous access. A disadvantage of a more lateral site is the decrease in diameter of the vein. The depth of the vein and lack of obvious surface landmarks suggest that ultrasound should be used for access. This route of access deserves wider study.


    References
 Top
 Abstract
 Introduction
 Surface landmark techniques
 Applied anatomy
 Methods
 Results
 Discussion
 Conclusion
 References
 
1 Nickalls RWD. A new percutaneous infraclavicular approach to the axillary vein. Anaesthesia 1987; 42: 151–4[ISI][Medline]

2 Taylor BL, Yellowlees I. Central venous cannulation using the infraclavicular axillary vein. Anesthesiology 1990; 72: 55–8[ISI][Medline]

3 Mogil RA, DeLaurentis DA, Rosemond GP. The infraclavicular venipuncture. Arch Surg 1967; 95: 320–4[ISI]

4 Browse NL, Burnand KG, Irvine AT, Wilson NM. Diseases of the Veins, 2nd Edn. London: Arnold, 1999; 45

5 Borja AR, Hinshaw JR. A safe way to perform infraclavicular subclavian vein catheterization. Surg Gynecol Obstet 1970; 130: 673–6[ISI][Medline]

6 Rominger CJ. The normal axillary venogram. Am J Roentgenol 1958; 80: 217–24

7 Yeow KM, Kaufman JA, Rieumont MJ, Geller SC, Waltman AC. Axillary vein puncture over the second rib. Am J Roentgenol 1998; 170: 924–6[ISI][Medline]

8 Hughes P, Scott C, Bodenham A. Ultrasonography of the femoral vessels in the groin: implications for vascular access. Anaesthesia 2000; 55: 1198–1202[CrossRef][ISI][Medline]

9 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[Abstract]

10 Kwon TH, Kim YL, Cho DK. Ultrasound-guided cannulation of the femoral vein for acute haemodialysis access. Nephrol Dial Transplant 1997; 12: 1009–12[Abstract]

11 Schregel W, Hoer H, Radtke J, Cunitz G: Ultrasonic guided cannulation of the axillary vein in intensive care patients. Anaesthesist 1994; 43: 674–9[CrossRef][ISI][Medline]

12 Greher M, Retzl G, Niel P, Kamolz L, Marhofer P, Kapral S. Ultrasonographic assessment of topographic anatomy in volunteers suggests a modification of the infraclavicular vertical brachial plexus block. Br J Anaesth 2002; 88: 632–6[Abstract/Free Full Text]

13 Mallory DL, Shawker T, Evans RG, et al. Effects of clinical maneuvers on sonographically determined internal jugular vein size during venous cannulation. Crit Care Med 1990; 18: 1269–73[ISI][Medline]

14 Mangar D, Turnage WS, Mohammed SA. Is the internal jugular vein cannulated during insertion or withdrawal of the needle during central venous cannulation? Anesth Analg 1993; 76: 1375

15 Marayuma K, Nakajima Y, Hayashi Y, Ohnishi Y, Kuro M. A guide to preventing deep insertion of the cannulating needle during catheterization of the internal jugular vein. J Cardiothorac Vasc Anesth 1997; 11: 192–4[CrossRef][ISI][Medline]

16 Ellison N, Jobes DR, Troianos CA. Internal jugular vein cannulation. Anesth Analg 1994; 78: 198