Share this article on:

Guiding Surgical Cannulation of the Inferior Vena Cava with Transesophageal Echocardiography

Kirkeby-Garstad, Idar MD*,; Tromsdal, Arve MD†,; Sellevold, Olav F. M. MD, PhD*,; Bjørngaard, Mads MD*,; Bjella, Lise K. MD*,; Berg, Einar M. MD*,; Karevold, Asbjørn MD‡,; Haaverstad, Rune MD, PhD‡,; Wahba, Alexander MD, PhD‡,; Tjomsland, Ole MD, PhD‡,; Astudillo, Rafael MD, PhD‡,; Krogstad, Arne CCP‡,; Stenseth, Roar MD, PhD*

doi: 10.1213/01.ANE.0000055361.50727.11

We studied 150 adult cardiac surgery patients to assess visualization of the venous cannula and the venous system by intraoperative transesophageal echocardiography and to register the incidence of cannulation of hepatic veins. The quality of images, the dimensions of the venous system, the position of the venous cannula, and the adequacy of venous return were registered. Acceptable image quality of the inferior vena cava and the right hepatic vein (RHV) was obtained in 95% and 87% of cases, respectively. Considerable individual variations were found in the dimensions of the venous system. The cannula position could be determined in 99% of the cases. Ten percent of venous cannulae were primarily placed in the RHV. A short distance between the eustachian valve and the RHV possibly predisposes to cannulation of the RHV. No other patient-related factors were associated with cannula position. Placement of the cannula deep in the inferior vena cava was associated with reduced venous return and may be a more important cause of reduced return than a cannula positioned in a hepatic vein.

Departments of *Anaesthesia,

†Cardiology, and

‡Surgery, St. Elisabeth Heart Centre, University Hospital of Trondheim, Norway

Presented in part at the annual meeting of the European Association of Cardiothoracic Anaesthesiologists, Dublin, Ireland, June 2002.

December 17, 2002.

Address correspondence and reprint requests to Idar Kirkeby-Garstad, MD, Department of Anesthesia and Intensive Care, St. Elisabeth Heart Centre, University Hospital of Trondheim, Hans Nissensgt 3, N 7018 Trondheim, Norway. Address e-mail to

During cardiopulmonary bypass (CPB), the venous return to the extracorporeal circuit depends on, among other factors, the position of the venous cannula (1). The intended position of the cannula is the inferior vena cava (IVC).

The use of intraoperative transesophageal echocardiography (TEE) is a standard procedure during cardiac operations in our institution. From September 2000 through the start of this study, we intermittently observed the venous cannula position by TEE. Approximately 10% of the venous cannulae were found to have been placed in the right hepatic vein (RHV). This prompted a prospective observational study to evaluate the value of intraoperative TEE in identifying the position of the venous cannula.

The primary aim of this study was to assess the degree and quality of visualization of the venous cannula and the venous system by intraoperative TEE during cardiac surgery and to define the incidence of cannulation of hepatic veins. Second, we investigated whether venous return was influenced by the anatomic position of the cannula.

Back to Top | Article Outline


The study protocol was approved by the Regional Board of Ethics in Medical Research. We planned to include a sufficient number of patients to determine the incidence of cannulation of hepatic veins. Consecutive patients undergoing cardiac surgery with the use of CPB were enrolled in the study. Written, informed consent was obtained. Because this was an observational study, patients with any contraindication toward TEE were not included.

Commercially available venous cannulae were used. Routinely, a two-stage cannula (Medtronic, Minneapolis, MN) sized 36/51F in men and 34/46F in women was inserted. The designation of a two-stage cannula refers to the size of the two segments of the cannula (Fig. 1). The large cannula types are designed for gravity-dependent (siphon) drainage. When vacuum-assisted drainage was intended (i.e., subatmospheric pressure in the venous reservoir augmenting the siphon effect), either a Medtronic 29/29F or a Baxter (Thousand Oaks, CA) 29/37F two-stage cannula was inserted. If separate cannulation of the superior vena cava and IVC was indicated, a straight single-stage Medtronic 28F or 32F cannula was used in the IVC. The two-stage cannula types were all inserted via the right atrial appendage, whereas the single-stage cannulae were inserted through the lower part of the right atrium. After the induction of anesthesia, air was evacuated from the stomach with a gastric tube before a TEE examination was performed with a GE Vingmed System Five or Vingmed CFM 750 (GE Vingmed, Norway) by using a multiplane probe.

Figure 1

Figure 1

The RHV and the middle and left hepatic veins all enter the IVC shortly before it empties into the right atrium. To identify the hepatic veins by TEE, we start at the level of the aortic valve at 0° (Fig. 2a). The probe is advanced and turned to the patient’s right (clockwise) to obtain a good image of the right atrium with the tricuspid valve and coronary sinus (Fig. 2b). The probe is further turned to the right and, if necessary, advanced to identify the orifice of the IVC (Fig. 2c) and to bring it to the center of the sector (Fig. 2d). The transducer plane is rotated to 40°–60°, keeping the IVC in focus (Fig. 2e), and the probe is subsequently advanced 2–3 cm along the IVC. To identify the RHV (Fig. 2f), it may be necessary to turn the probe to the right. Because of variations in anatomy, this procedure must be modified and individually adapted to the actual patient. The anatomy of the hepatic veins is outlined in Figure 3. The middle and left hepatic veins are found at the same esophageal level as the RHV. Starting from the RHV, the middle vein is found by rotating the transducer to 50°–90° and turning the probe counterclockwise back toward the normal position. To identify the left hepatic vein, the probe is turned further counterclockwise, with the transducer rotated to 80°–130°. In the images acquired, the middle hepatic vein runs in the same direction as the RHV, whereas the left hepatic vein runs more transversally to the echo beam.

Figure 2

Figure 2

Figure 3

Figure 3

Cannulae unintentionally placed in a hepatic vein will, because of the anatomical site, most likely end up in the RHV. It is the largest and enters the IVC at an oblique angle. The venous cannula is furthermore directed slightly toward the patient’s right during insertion. We therefore focused on identification of the IVC and the RHV.

Visualization of the IVC and the RHV was performed before cannulation and was evaluated separately. We graded visualization according to the following scale: “good” (a clear view of the vessel and its margins), “acceptable” (the course of the vessel could readily be identified), “poor” (the course of the vessel could be identified with difficulty), or “not seen.”

To prepare mapping of the cannula position relative to the IVC and RHV, the distance from the eustachian valve (E) to the RHV and the diameters of the IVC and the RHV were measured. The upper and lower margins of the orifice of the RHV were identified and named t and s, respectively (Fig. 3). The E-t and E-s distances were measured. The diameter of the IVC was measured at Point s (Fig. 3), and the diameter of the RHV was measured 1 cm distal to Point s (Fig. 3).

The IVC and the RHV were monitored during cannulation to determine the placement of the cannula. For cannulae placed in the IVC, the distance from the E valve to the tip was estimated. Visualizing both the E valve and the end of the cannula in one picture is often not possible, and the tip itself is not always easily identified. To obtain the s-tip measurement, we measured the distance from Point s (Fig. 3) to the end of the metal coil in the cannula wall and added the known distance from the end of the coil to the cannula tip. The distance E-tip was calculated by adding the E-s and s-tip measurements. The central venous pressure (CVP) before cannulation and after final placement of the cannula was obtained from a catheter in the superior vena cava.

Six experienced anesthesiologists and one cardiologist performed the TEE examinations. The finding of a cannula positioned outside the IVC was communicated to the surgical team before CPB was started. The repositioning of the cannula was left to the surgeon to decide, and the number of attempts needed to obtain a position in the IVC was registered. After stable CPB was obtained, the venous return was clinically evaluated by the perfusionist and classified as “good,” “acceptable,” or “poor.” All data were obtained during surgery, and a registration form was completed by the investigator. Most data were specific to the study and not registered otherwise. Incomplete registrations were treated as missing values.

Normally distributed data are presented as mean ± sd. Sex differences in dimensions of the venous system were explored with Student’s t-test. Logistic regression was used to analyze the effect of patient-related factors on the incidence of cannulae placed in the RHV. Differences in frequency of cannulation of the RHV between cannula types were explored by an extended Fisher’s exact test. The uneven distribution of the cannulae used did not allow further post hoc tests.

Back to Top | Article Outline


The study was conducted between December 2001 and March 2002. We did not find any difference regarding cannulation of hepatic veins among Patients 1–50, 51–100, or 101–150. The study was therefore stopped after inclusion of 150 patients. The patients were 37 women and 113 men; mean age was 66 ± 9 yr, height was 173 ± 8 cm, and weight was 80 ± 14 kg. Coronary artery bypass grafting was performed in 112 patients, aortic valve replacement in 14, combined aortic valve replacement and coronary artery bypass grafting in 20, and other surgical procedures in 4. There were two reoperative procedures. In six cases the heart was adherent to the surrounding tissues. Surgical access to the right atrium was extremely difficult in one case. No patient had a permanent pacemaker or other devices that interfered with access to the right side of the heart.

The visualization of the IVC was registered as good or acceptable in 95% of cases and of the RHV in 87% of cases (Table 1). The mean distances measured during the mapping of the venous system are shown in Table 2. Considerable variations were observed for all these variables. The E-s distance was significantly larger in men than women (P = 0.04), but no other sex differences were found.

Table 1

Table 1

Table 2

Table 2

On two occasions, difficulties positioning the cannula into the IVC were encountered. In both cases, TEE revealed that a long, floppy E valve was the probable cause. One of the cannulae was correctly placed after four attempts, and in the other case, a basket cannula was used in the right atrium.

TEE enabled determination of the cannula position in 148 (98.7%) of 150 cases. In a 49-yr-old male patient (height, 173 cm; weight, 100 kg) with an enlarged heart, neither the IVC/RHV nor the cannula could be visualized. The surgical procedure revealed a rotated heart with a posterior position of the right atrium. Displacement of the mediastinal structures was the probable cause of missed visualization. In the other case in which the position of the cannula could not be determined, visualization of the IVC and RHV was poor before cannulation.

Control of cannula position by TEE demonstrated that 14 cannulae (9.5%) were placed outside the IVC. Twelve had the tip in the RHV, one was placed in a second RHV, and one was coiled in the right atrium.

A significant correlation was found between the E-t distance and cannulation of a hepatic vein (r2 = 0.103;P = 0.03). A short E-t distance increased the incidence of such cannulation. None of the other measured dimensions of the venous system influenced placement. The apparent sex difference in cannulae placed outside the IVC (women: 6 of 37, 16.2%; men: 8 of 110, 7.3%) did not reach statistical significance (P = 0.11). Neither body length (P = 0.24) nor body mass index (P = 0.99) was associated with the incidence of cannulae positioned outside the IVC.

Placement in the hepatic veins was significantly influenced by cannula type (P < 0.001). The uneven distribution of cannulae prevented further statistical analysis of cannula type. The frequencies of placement outside the IVC for different cannula types are given in Table 3.

Table 3

Table 3

All but one cannula placed in the IVC had the tip distal to the inlet of the hepatic vein. The mean distance from the E valve to the tip of the cannula was 8.4 ± 1.9 cm (range, 1.0–12.5 cm) (n = 109).

The CVP was 5.8 ± 3.3 mm Hg before cannulation versus 4.5 ± 3.8 mm Hg after established CPB when the cannula was placed <10 cm into the IVC; it was 5.1 ± 2.2 mm Hg versus 4.6 ± 1.9 mm Hg with the cannula placed >10 cm into the IVC (not significant). When the cannula was located in the RHV, the CVP was 4.6 ± 3.1 mm Hg versus 3.6 ± 2.5 mm Hg, respectively (not significant).

From the start of the study, all venous cannulae that were located outside the IVC were repositioned before CPB was initiated. Eight cannulae placed in the hepatic veins and one placed in the right atrium required an average of 1.5 ± 0.8 (range, 1–3) attempts before they were positioned in the IVC. The five last cannulae observed in the RHV were, however, not repositioned. In each case, the surgeon decided to initiate bypass and observe whether return was sufficient before eventual repositioning of the cannula. Venous return was characterized as “good” in four cases and “acceptable” in one case.

Venous return was registered in 143 patients. It was characterized as “good” in 113 and “acceptable” in 22 patients, giving good or acceptable venous return in 94% of the cases. Poor venous return was observed in eight cases. A closer analysis of these cases revealed that the cannula was placed deep into the IVC in five patients (mean E-tip distance, 10.3 ± 1.3 cm) and in a “normal” position in one patient (E-tip distance, 6.2 cm). In the last two cases, the cannula was repositioned before any measurements could be made. No increase in CVP was observed in any of these patients after CPB was initiated.

Back to Top | Article Outline


This study shows that TEE allows satisfactory determination of the position of the venous cannula in relation to the IVC and RHV in nearly all patients subjected to CPB. The incidence of venous cannulae primarily placed in the RHV was nearly 10% (Fig. 1). The surgeons, being aware of the continuing study, may have reduced the actual numbers of RHV cannulations in the study period. No cannula was positioned in the middle or left hepatic vein. The three main hepatic veins usually enter the IVC at almost the same distance from the right atrium. Consequently, a cannula seen in the IVC distal to the inlet of the RHV is unlikely to be positioned in one of the other main hepatic veins. Monitoring the IVC and RHV will therefore detect whether a hepatic vein is cannulated.

We produced good- or acceptable-quality images of the IVC and the RHV in 95% and 87% of cases, respectively. Meierhenrich et al. (2) studied 34 patients scheduled for abdominal surgery and identified the three main hepatic veins by multiplane TEE in all patients. Pinto et al. (3), Nomura et al. (4), and Gårdebäck et al. (5), studying 29, 45, and 8 cardiac surgical patients, respectively, obtained hepatic vein flow patterns by TEE in all cases before cannulation. These studies all support the position that the IVC and RHV are well visualized by TEE.

The cannulae used in this study contain a metal coil that may induce acoustic shadowing and obscure the image of the RHV and the IVC. Furthermore, the hepatic veins collapse to some degree during CPB circulation. Imaging of the IVC and hepatic veins is therefore easier before cannulation and initiation of CPB. Identification of the central veins before and close monitoring of them during cannulation makes it easier to decide whether the cannula is correctly positioned.

A short distance between the E valve and the upper margin of the inlet of the RHV (E-t) was associated with a significantly increased incidence in cannulation of the hepatic veins. The distance between the E valve and the lower margin of the same inlet (E-s) had no influence on cannulation of the hepatic veins. These measurements are not truly independent, and hence this result should be interpreted with caution.

Statistically, we cannot identify which cannula type is more likely to enter the hepatic vein, despite the uneven frequency displayed in Table 3. The small-diameter cannula (29/29F) was not more frequently placed in hepatic veins. This indicates that the cannula diameter is not decisive for positioning in a hepatic vein. The uneven tendency to enter the hepatic veins may depend on the difference in the site of insertion or the difference in the frequency of use.

The physiological consequences of cannulae placed in the hepatic veins were not the primary focus of our study. The surgeons were therefore informed of the finding, and most cannulae in the hepatic veins were immediately repositioned. Acceptable flow from five cannulae in the hepatic vein showed that venous return was not necessarily impeded by hepatic vein placement. This corresponds well with the fact that the observed frequency of cannulae placed in hepatic veins exceeds the generally observed frequency of problems with venous return. The perfusionists evaluated venous return after being informed of the position of the cannula. This may have influenced their evaluation of venous return to some degree. Their estimation of poor venous return was, however, based on the specific criteria of having sufficient venous return to meet a preoperative set pump flow.

The TEE mapping of the venous system should not be considered as anatomically exact measurements. The IVC and RHV diameters were taken from a longitudinal view and may consequently underestimate the true values. The imaging plane may have been oblique, and E-t and E-s do not necessarily represent the shortest distance between the E valve and the RHV. These factors may have contributed to the considerable individual variation found for all these variables. The significantly larger E-s distance in men than in women probably reflects the larger height in men.

The E-tip distance is less accurate than the above-mentioned measurements because it was calculated from two measurements: E-s and s-tip. However, it still gives a good picture of cannula position in the venous system. Our results indicate that placement of the cannula deep in the IVC is associated with reduced venous return.

A casuistic report has recently been published proposing TEE as a potential method to determine the position of venous cannulae in cardiac surgery (6). However, no systematic TEE study of surgical cannulation of the IVC has previously been published.

As TEE has become an integrated part of cardiac anesthesia and surgery, the extra effort of examining cannula placement is negligible. Placement of the cannula in a hepatic vein is frequent and easily detected, especially if the IVC/RHV is monitored during cannulation. A position of the cannula deep in the IVC is also easily visualized. Further studies are needed before the full clinical usefulness of this method can be determined. We strongly recommend that the position of the venous cannula be determined in all cases of poor venous return during CPB.

Back to Top | Article Outline


1. Bennett EV Jr, Fewel JG, Ybarra J, et al. Comparison of flow differences among venous cannulas. Ann Thorac Surg 1983; 36: 59–65.
2. Meierhenrich R, Gauss A, Georgieff M, Schütz W. Use of multi-plane transoesophageal echocardiography in visualization of the main hepatic veins and acquisition of Doppler sonography curves: comparison with the transabdominal approach. Br J Anaesth 2001; 87: 711–7.
3. Pinto FJ, Wranne B, St. Goar FG, et al. Hepatic venous flow assessed by transesophageal echocardiography. J Am Coll Cardiol 1991; 17: 1493–8.
4. Nomura T, Lebowitz L, Koide Y, et al. Evaluation of hepatic venous flow using transesophageal echocardiography in coronary artery bypass surgery: an index of right ventricular function. J Cardiothorac Vasc Anesth 1995; 9: 9–17.
5. Gårdebäck M, Settergren G, Brodin L-Å. Hepatic blood flow and right ventricular function during cardiac surgery assessed by transesophageal echocardiography. J Cardiothorac Vasc Anesth 1996; 10: 318–22.
6. Sastre Rincón JA, Hernández Valero A, Fernández Pérez A, Muriel Villoria C. A simple method to control a malposition of inferior venous cannula in cardiac surgery using transesophageal echocardiography. Eur J Cardiothorac Surg 2002; 22: 146.
© 2003 International Anesthesia Research Society