Secondary Logo

Journal Logo

Original Article

ECG-guided central venous catheter positioning: does it detect the pericardial reflection rather than the right atrium?

Schummer, W.*; Schummer, C.*; Müller, A.; Steenbeck, J.; Fuchs, J.*; Bredle, D.; Hüttemann, E.*

Author Information
European Journal of Anaesthesiology: August 2004 - Volume 21 - Issue 8 - p 600-605


Central venous catheters (CVCs) have become indispensable in medical practice. About 200 000 of them are placed annually in the UK [1,2]. At our 1350 bed University Hospital, about 7500 CVCs are inserted each year. Central venous cannulation puts patients at increased risk for mechanical and infectious complications, some of which are potentially fatal. Certain complications, such as vessel perforation, are related to the position of the tip and the insertion site. However, any position of the tip can result in serious complications. Furthermore, there is a lack of good evidence on which to base practice of the catheter tip position [3].

In order to decrease the risk of superior vena cava (SVC) perforation - one of the most dangerous complications of central venous cannulation [4] - the catheter tip should be placed parallel with the long axis of the SVC, such that the tip does not abut the wall of the vein or heart. In 2000, Schuster and colleagues published an anatomical study concluding that CVC tips should be located in the SVC above the level of the carina, in order to avoid cardiac tamponade [5]. The carina is seen as a radiological landmark located outside the pericardial sac. Currently, electrocardiography (ECG)-guided CVC placement is accepted as a useful and safe method for correct catheter tip positioning [4].

In a previous study on ECG guidance, many triplelumen catheters inserted through the left internal jugular vein (LIJV) had to be advanced further than indicated by ECG guidance in order to achieve free venous back flow through all lumina [6]. In contrast to what has been expected, more than 90% of these catheter tips were still located outside the right atrium (RA). We therefore concluded that the appearance of an 'intra-atrial' ECG with its increase in P-wave amplitude corresponds to a structure outside the RA. We conducted this prospective study comparing CVC positioning via the right internal jugular vein (RIJV) (Group R) and LIJV (Group L) to give clinical evidence for this hypothesis.


The study protocol was approved by our Hospital Ethics Committee. This prospective, randomized, single centre study was performed in the division of cardiothoracic and vascular anaesthesia of a university hospital. Patients scheduled for cardiac surgery were eligible for the study, which ran from January to April 2003. For this type of surgery, CVC placement and transoesophageal echocardiography (TOE) monitoring is routine in our department. Postoperatively, all patients were admitted to the intensive care unit (ICU) and a chest radiograph was taken. Exclusion criteria were cardiac rhythms other than sinus rhythm after induction of anaesthesia, contraindication to TOE (e.g. gastric or oesophageal pathology or surgery). Patient characteristics recorded were age, gender, height and weight, as well as the type of surgery (Table 1). All CVCs were placed by one anaesthetist (W.S.).

Table 1
Table 1:
Patient characteristics of the two groups: RIJV and LIJV approach.

For blockwise randomization, computer-generated numbers were used. The patients were assigned to the RIJV (Group R) or LIJV approach (Group L). After induction of general anaesthesia, patients were placed in a 20° Trendelenburg position for CVC insertion. The RIJV or LIJV, respectively, was punctured midway between the mastoid process and the sternal notch, just lateral to the pulsation of the carotid artery. First, the vein was entered using a sterile Seldinger technique. Then the guide-wire was advanced and a triple-lumen polyurethane CVC introduced (Certofix® Trio SB 730, length 30 cm, 7 French). The distances of the opening ports are 2.5 and 5 cm from the distal opening for the middle and proximal lumen, respectively. Markings on the catheter with a distance of 1 cm allowed the measurement of its depth of insertion (Fig. 1). The kit includes a connection cable with a crocodile clip for connecting the guide-wire to a Certodyn® universal adapter (both: B. Braun Melsungen AG, Melsungen, Germany).

Figure 1
Figure 1:
Triple-lumen CVC. (A) Guide-wire protruding the distal port; (B) port opening of the middle lumen in 2.5 cm distance to the catheter tip; (C) guide-wire inserted through hub of distal lumen. Black marking with crocodile clamp.

In each patient, two methods for positioning the catheter tip were sequentially applied: ECG guidance using the Seldinger guide-wire in the distal lumen (Method A), and ECG guidance using a hypertonic saline fluid column in the middle lumen of the catheter (Method B). Free venous back flow from each port of the catheter was tested before it was fixed by a suture to the skin. The final position of all CVC tips was checked by TOE and postoperatively by portable chest radiography.

Method A

First, the guide-wire was used as a unipolar electrode. A black marking on the proximal end of the guide-wire indicates the point at which the tip of the wire levels with the port of the distal catheter lumen (Fig. 1). A sterile connection cable was clamped to the guide-wire at the marked position in order to connect it with an adapter that allows the operator to switch from a surface (Einthoven Lead II) to an intravascular ECG. The catheter was advanced together with the guide-wire until an increase in P-wave size was detected. Then they both were withdrawn stepwise until the P-wave returned to its normal size. The insertion depth, the ID (ID-guide-wire) was recorded in whole centimetres.

Method B

The middle lumen filled with saline 10% served as detecting electrode. Alphacard® (B. Braun, Melsungen, Germany), an electrical-conductive syringe with a cable and a cable joint for connection to the Certodyn® universal adapter was used [7]. With the guide-wire in place the catheter was advanced until the first P-wave increase was noted. The guide-wire was removed, than all lumina were tested for free venous backflow. The ID was recorded again (ID-saline) and the catheter was sutured to the skin (Fig. 2).

Figure 2
Figure 2:
Flow chart of the study design.

After placing and testing the CVC, the TOE probe (multiplane probe, 6.2 MHz/HP Sonos 5500®, Philips, Andover, USA) was inserted to a midoesophageal position and rotated to the right (clock-wise). Then the plane of the probe was turned to an angle of 90-110° to produce a bicaval view according to the American Society of Echocardiography/Society of Cardiovascular Anesthesiologists (ASE/SCA) guidelines [8]. The echocardiographic correlate of the SVC-RA junction was defined as the base of the superior edge of the crista terminalis [9,10]. A rapid flush of cephazolin 2 g in 20 mL physiological saline (given for perioperative antibiotic prophylaxis, injected through the distal lumen) was used to identify the tip of the CVC by TOE (Fig. 3). The microbubbles of the solution acted as contrast medium and helped to identify the plane of the SVC in which the catheter tip was positioned. The catheter tip was usually identified as two closely spaced, parallel, bright echo dense lines surrounding the darker fluid-filled lumen. The relationship of the CVC to the crista terminalis was recorded as a benchmark for the ID. All TOE examinations were performed by a second anaesthetist who was blinded to the IDs of the CVCs.

Figure 3
Figure 3:
The echocardiographic SVC-RA junction was defined as the base of the superior edge of the crista terminalis (*). The catheter tip was usually identified as two closely spaced, parallel, bright echo dense lines surrounding the darker fluid-filled lumen. Rapid flush of cephazolin 2 g in physiological saline 20 mL was used to identify the distal end of the CVC by TOE (arrow). The microbubbles of the solution act as contrast medium.

Within 3 h after surgery, a portable chest radiograph was taken. The radiographic SVC-RA junction was defined as the apex of the concave shadow formed by the superimposition of the distal SVC on the RA [11,12]. The catheter position was determined by a radiologist, also blinded to the study on CVC insertion. A correct position of the catheter tip was defined as to be in the SVC or at the SVC-RA junction and parallel (<40°) to the SVC wall, according to chest radiography and TOE. Malposition was considered to be a catheter tip placed in a vein other than the SVC, or in the RA or if the tip abutted the vein or heart wall at an angle exceeding 40°. Figure 4 depicts a correctly positioned catheter while Figure 5 is an example of a poorly positioned catheter (not seen in this study). The patient archiving and communication system (PACS) (Image Devices GmbH, Idstein, Germany) was used to view the digital images of the chest radiograms. The angle between the distal catheter and the lateral wall of the SVC was quantified using the viewing software of the PACS.

Figure 4
Figure 4:
Chest radiograph (CXR) with correctly positioned CVC. Catheter tip marked with an arrow.
Figure 5
Figure 5:
CXR with incorrectly positioned CVC. The catheter tip (arrow) abuts the wall of the cava superior at a steep angle of approximately 90°.

Data analysis

The descriptive data are expressed as mean ± standard deviation (SD). Data analysis was performed using SPSS® 11.0 (SPSS Inc., Chicago, IL, USA) under Windows XP® (Microsoft Corp., Redmond, CA, USA).


Of the 101 patients entering the study, all but one could be analysed. One patient could not be analysed because of inadvertent CVC withdrawal during transport to the ICU. Fifty central venous catheters were inserted through the RIJV (Group R), and 51 through the LIJV (Group L). The depth of insertion of 50 catheter tips in Group R and 50 in Group L were analysed. The mean depth of insertion in Group R was 15.7 cm (SD = 1.5) with Method A (ID-guide-wire), and 17.6 cm (SD = 1.5) with Method B (ID-saline), the mean difference was 2.0 cm (SD = 0.2). In Group L, the mean depth of insertion with Method A (ID-guide-wire) was 20.4 cm (SD = 2.2), and 22.4 cm (SD = 2.2) with Method B (ID-saline), the mean difference was 2.0 cm (SD = 0.3).

All CVCs had free venous back flow through all lumina with ID-saline. There were no malpositions into vessels other than the SVC. Using TOE, no CVC was found to have entered the RA. Using chest radiography, no catheter was seen abutting the lateral wall of the SVC (Table 2).

Table 2
Table 2:
Results of the two groups: RIJV and LIJV approach.


ECG guidance has become accepted as a tool to help in the placement of the tip of a CVC.

In many hospitals, ECG guidance for CVC placement has replaced a post-insertion chest radiography to check the position. From a legal viewpoint, in Germany at least, ECG guidance is equivalent to a chest radiograph [13] - the situation may be different in other countries. Unless the subclavian vein is punctured, or the procedure is accompanied by some difficulties, post-insertion chest radiography is considered superfluous if the CVC is placed with ECG guidance.

Proper positioning of CVC concerns two major issues: the concern for patient safety versus the desire for optimal catheter performance. There are numerous reports describing complications attributed to central venous catheter tips positioned within the RA. Although most cases of catheter perforation occur in the RA, perforation of the SVC can also occur [14]. If SVC perforation occurs within the pericardial reflection, a potentially life-threatening cardiac tamponade can still occur. Regarding the pericardial reflection, the major problem is the lack of reliable surface land-marks. Furthermore it cannot be identified on chest radiography.

In vitro and in vivo studies have shown that an acute angle of the catheter of more than 40° to the wall of the SVC imposes a markedly increased risk of vessel perforation [15]. The risk of perforation is negligible if the catheter runs parallel with the vein. From a clinical point of view, there is no question that a CVC tip should not abut a vessel wall. The significance of this problem increases with the stiffness of the catheter (e.g. multiple-lumen catheters, haemodialysis catheters) [16-20]. In CVCs inserted through the RIJV, abutting of the tip against the vessel wall does only seldom occur. Due to the straight course of a CVC, insertion is best via the RIJV in the neck and upper thorax. Compared with the right side, a catheter inserted through the LIJV is distorted by two curves of up to 90° (the left internal jugular into the left innominate vein and the left innominate vein into the SVC). In this situation the benefit of keeping the catheter tip above the pericardial reflection (thus preventing pericardial tamponade) may be offset by the risk of having the catheter abut the SVC at an acute angle (thus possibly perforating the SVC). Current guidelines on CVC tip position focus only on the risk of tamponade [21]. We believe that an unsatisfactory tip position outside the heart should not be accepted purely to satisfy these guidelines [3].

ECG-guided CVC placement claims to guide the intra-atrial position of the CVC tip by detecting an intra-atrial P-wave (P atriale) with the exploring electrode. The adjacent atrial wall tissue is thought to be responsible for the increase in voltage of the P atriale [3,7]. Based on that assumption the manufacturer recommends withdrawing the catheter by 1-2 cm after normalization of P-wave amplitude to ensure placement outside the atrium: The mean distance between the two catheter ports of the catheter used in our study was 2.5 cm. Provided both techniques identified the same structure, an increase of ID of approximately 2.5 cm between Methods A and B was mandatory.

In our study ECG guidance by saline 10% in the middle lumen, resulted in a mean advancement of 2.0 cm after primary ECG positioning with a guidewire in the distal catheter lumen in all catheters (this might be explained by the gross scale unit on the catheter and the oblique course of the pericardial reflection). We therefore conclude that we always detected the same structure by both methods of ECG guidance. TOE served as the gold standard in assessing the final CVC position with respect to the RA. In all patients the tip of the CVC could be localized by TOE. The multiplane probe employed allowed a three dimensional assessment of the position of the catheter in relation to the RA and SVC without change of the transducer position. No catheters entered the RA.

Regarding our methodology of positioning the catheter tip 2.0 cm beyond the point where the P-wave had just returned to normal size, one would have expected all 100 CVC tips to be placed within the upper RA. But in our study, not a single CVC was seen beyond the base of the crista terminalis in the RA.

According to this data, we conclude that the anatomical structure corresponding to the 'intra-atrial' ECG is not the RA but must be located higher up in the SVC. As the pericardium is a fluid-filled sac (important for electrical conductance), and there is no other anatomical structure beyond it to explain our findings, we conclude that the pericardial reflection is responsible for the electrical phenomenon of first P-wave increase. We agree with the opinion that a catheter withdrawn to a position where no 'intra-atrial' ECG can be recorded is located outside the heart [7]. However, we are convinced that catheters placed under such ECG guidance actually are located just outside of the pericardium. Thus a post-insertion chest radio-gram is still mandatory in left-sided CVCs placed by ECG guidance using the traditional manufacturers' guidelines to detect catheters placed in an acute angle. Our data are in contrast to the statement by McGee, that a CVC should not routinely be inserted to a depth of >20 cm [7]. In left-sided CVCs, abutting of the vessel wall can be avoided by a deeper insertion. Catheters of 20 cm length might be too short. A limitation of our study is the time delay between CVC insertion, TOE assessment and the final position check by postoperative chest radiography. Catheter migration due to surgical manipulations cannot be excluded.

We conclude that the structure detected by an increase in the P-wave is not the junction between the SVC and RA, but more likely the pericardial reflection point. In our study, placing CVC tip 2 cm beyond the structure detected by guide-wire ECG guidance through the distal lumen always resulted in reliably functioning catheters with their tips outside the RA. A modification of the standard approach to ECG-guided CVC placement, as we present here, would seem prudent, pending further supportive data. Anatomical studies to verify our thesis are currently underway.


1. Waghorn DJ. Intravascular device-associated systemic infections: a 2 year analysis of cases in a district general hospital. J Hosp Infect 1994; 28: 91-101.
2. Collier PE, Blocker SH, Graff DM, Doyle P. Cardiac tamponade from central venous catheters. Am J Surg 1998; 176: 212-214.
3. Fletcher SJ, Bodenham AR. Safe placement of central venous catheters: where should the tip of the catheter lie? Br J Anaesth 2000; 85: 188-191.
4. Scott WL. Complications associated with central venous catheters. A survey. Chest 1988; 94: 1221-1224.
5. Schuster M, Nave H, Piepenbrock S, Pabst R, Panning B. The carina as a landmark in central venous catheter placement. Br J Anaesth 2000; 85: 192-194.
6. Schummer W, Herrmann S, Schummer C, et al. Intra-atrial ECG is not a reliable method for positioning left internal jugular vein catheters. Br J Anaesth 2003; 91: 481-486.
7. McGee WT, Ackerman BL, Rouben LR, Prasad VM, Bandi V, Mallory DL. Accurate placement of central venous catheters: a prospective, randomized, multicenter trial. Crit Care Med 1993; 21: 1118-1123.
8. Shanewise JS, Cheung AT, Aronson S, et al. ASE/SCA guidelines for performing a comprehensive intraoperative multiplane transesophageal echocardiography examination: recommendations of the American Society of Echocardiography Council for Intraoperative Echocardiography and the Society of Cardiovascular Anesthesiologists Task Force for Certification in Perioperative Transesophageal Echocardiography. Anesth Analg 1999; 89: 870-884.
9. Nanda NC, Domanski HJ. Normal anatomy. Atlas of Transoesophageal Echocardiography. Baltimore, USA: Lippincott Williams & Wilkins, 1998: 1-5.
10. Nanda NC, Domanski HJ. Superior vena cava and right atrium. Atlas of Transoesophageal Echocardiography. Baltimore, USA: Lippincott Williams & Wilkins, 1998: 36-38.
11. Wunderbaldinger P. "Tubes and lines' in thoracic radiography: malpositioning and complications. Wien Med Wochenschr 2001; 151: 524-531.
12. Trigaux JP, Goncette L, Van Beers B, de Wispelaere JF, Pringot J. Radiologic findings of normal and compromised thoracic venous catheters. J Thorac Imaging 1994; 9: 246-254.
13. Weissauer W. The cava-catheter from the medico-legal viewpoint. Anaesthesiol Intensivmed Notfallmed Schmerzther 1998; 33: 117-118.
14. Jay AW, Aldridge HE. Perforation of the heart or vena cava by central venous catheters inserted for monitoring or infusion therapy. Can Med Assoc J 1986; 135: 1143-1144.
15. Blackshear RH, Gravenstein N. Critical angle of incidence for delayed vessel perforation by central venous catheter: a study of in vitro data. Ann Emerg Med 1991; 21: 659.
16. NKF-DOQI clinical practice guidelines for vascular access. National Kidney Foundation-Dialysis Outcomes Quality Initiative. Am J Kidney Dis 1997; 30: S150-S191.
17. Barton BR, Hermann G, Weil III R. Cardiothoracic emergencies associated with subclavian hemodialysis catheters. JAMA 1983; 250: 2660-2662.
18. Kappes S, Towne J, Adams M, Kauffman HM, Maierhofer W. Perforation of the superior vena cava. A complication of subclavian dialysis. JAMA 1983; 249: 2232-2233.
19. Kielstein JT, Abou-Rebyeh F, Hafer C, Haller H, Fliser D. Right-sided chest pain at the onset of haemodialysis. Nephrol Dial Transplant 2001; 16: 1493-1495.
20. Twardowski ZJ. Advantages and limits of the jugular catheter approach. Nephrol Dial Transplant 1995; 10: 2178-2182.
21. Food and Drug Administration. Precautions necessary with central venous catheters. FDA Task Force. FDA Drug Bull 1989; July: 15-16.


© 2004 European Academy of Anaesthesiology