Secondary Logo

Journal Logo

The Use of Transesophageal Echocardiography for Preload Assessment in Critically Ill Patients

Tousignant, Claude P. MD, FRCPC; Walsh, Fergus MB, MRCPI, FFARCSI; Mazer, C. David MD, FRCPC

doi: 10.1213/00000539-200002000-00021

IV volume is often administered to patients in an intensive care unit (ICU) to improve cardiovascular function. We investigated the relationship between stroke volume (SV) and left ventricular (LV) size by using transesophageal echocardiography (TEE) in a population of 20 ICU patients and 21 postoperative cardiac surgical patients. We also examined whether LV end diastolic area (EDA), by TEE, could identify patients who increased SV by 20% or more (responders) after 500 mL of pentastarch administration. There was only a modest relationship (r = 0.60) between the EDA and the SV in all patients. No relationship could be found between the pulmonary capillary wedge pressure (PCWP) and the EDA in all patients. Both responder and nonresponder PCWP increased significantly after volume administration. Only responder EDA increased significantly after volume administration. Responders had significantly lower EDA (15.3 ± 5.4 cm2) and PCWP (12.2 ± 2.2 mm Hg) when compared with nonresponders (20.2 ± 4.8 cm2) and 15.9 ± 3.1 mm Hg, respectively). Few ICU patients and only those with a small EDA responded to volume administration. It was not possible to identify an overall optimal LV EDA below which most patients demonstrate volume-recruitable increases in SV.

Implications In a ventilated intensive care unit and cardiac surgical population, transesophageal echocardiography and pulmonary artery catheter are sensitive in detecting changes in preload after volume administration. Few patients demonstrate volume-recruitable increases in stroke volume when compared to cardiac surgical patients. It is not possible to establish an overall end diastolic threshold below which a large proportion of ventilated patients respond to volume administration.

Department of Anaesthesia, St. Michael’s Hospital, University of Toronto, Toronoto, Ontario, Canada

October 28, 1999.

This study was funded in part by the Zeneca Pharma Inc. Canadian Research Award in Anaesthesia awarded by the Canadian Anaesthetists’ Society to CPT.

Address correspondence and reprint requests to Dr. Claude Tousignant, Department of Anaesthesia, St. Michael’s Hospital, 30 Bond St., Toronto, Ontario, Canada, M5B 1W8. Address e-mail to

Optimizing patient hemodynamics in the intensive care unit (ICU) can be challenging. Patients often present with varied and severe hemodynamic disturbances. The pulmonary artery catheter has been used to aid in determining preload, afterload and myocardial performance. The values obtained however, can in some circumstances be misleading. The use of transesophageal echocardiography (TEE) in the critical care setting has been increasing (1–6). TEE provides rapid visualization of left ventricular (LV) dimensions and function. The LV end diastolic area (EDA) in the transgastric midpapillary short axis view has been used as a measure of preload in vascular and cardiac surgical patients (7–11). It has also been assessed in critically ill patients (12–15). Our goal was to determine the relationship between stroke volume (SV) and EDA in an ICU population and whether an optimal EDA exists below which an increase in SV is observed after IV volume administration. We then examined whether an ICU patient subgroup was different from postoperative cardiac surgical patients.

Back to Top | Article Outline


After approval by the hospital ethics review board, written, informed consent was obtained from the patients or their families. There were two groups of patients. The ICU group consisted of all patients admitted to the ICU who did not undergo cardiac surgery. In the ICU group, 20 patients who the attending physician believed could benefit from IV volume administration (pulmonary capillary wedge pressure [PCWP] ≤ 20, inotropic support, low urine output) with adequate gas exchange were enrolled. The attending physician was not involved in the study. The cardiac surgical group consisted of 21 elective postoperative patients admitted to the cardiovascular unit after uncomplicated coronary artery bypass grafting. All patients had preoperative ejection fraction of 35% or more, had no recent myocardial infarction, and were in stable condition. Cardiac surgical patients were included if PCWP ≤ 20 and if there was adequate gas exchange. All TEE examinations were performed in sedated and ventilated patients with pulmonary artery catheters and radial arterial lines in situ. If inotropes were being administered, no changes in infusion rates were made during the study. Patients were excluded if there were contraindications to TEE or if there was presence of more than mild valvular disease.

Baseline measurements included heart rate, PCWP, and thermodilution cardiac output (CO). COs were taken as the average of three measurements using 10 mL of room temperature 5% dextrose solution. The SV was calculated as the CO divided by the heart rate.

By using an HP Sonos 2000 with omniplane probe (Hewlett Packard, Andover MA), the following baseline measurements were performed: in the transgastric midpapillary short axis view, the EDA and the end systolic area were measured by tracing the endocardial border including the papillary muscles. An average of three consecutive normal beats throughout the respiratory cycle was used for measurements. The fractional area of contraction, an estimate of ejection fraction, was calculated from the following relationship: (EDA − end systolic area)/EDA expressed in percentage. Patients with LV wall motion abnormalities were not excluded from this study. All TEE measurements were performed off line from tape recorded data by one observer. The TEE observer was not blinded to the results; however, the hemody- namic data were not available during off line TEE measurements.

After baseline measurements, 500 mL of pentastarch was administered IV over 15 min. Hemodynamic and TEE measurements were then repeated after this fluid challenge. Responders to the fluid challenge were defined as those patients who demonstrated an increase in SV of 20% or more.

Linear regression was used to examine the relationship between EDA and PCWP. Linear regression was also used to examine the relationship between SV and EDA as well as PCWP. Student’s t-test was used to compare the baseline hemodynamic and TEE data between the two groups. ANOVA with Bonferroni t-test was used to compare pre- and postbolus PCWP and EDA and to compare prebolus EDA and PCWP between both responders and nonresponder groups.

Back to Top | Article Outline


Forty-one patients were enrolled in this study, 20 in the ICU group and 21 in the cardiac surgical group. There was no significant difference with respect to age, height and weight between the groups. The ICU group consisted of 9 men and 11 women, whereas there were no women in the cardiac surgical group. Hemodynamic data for both groups are demonstrated in Tables 1 and 2. Patient 7 (Table 2) in the cardiac surgical group was excluded from pentastarch administration as his baseline PCWP was > 20 accompanied by a large EDA (39.5 cm2). Admitting diagnoses for the ICU group included sepsis (n = 7), ischemic heart disease (n = 3), cardiac arrest (n = 2), pancreatitis (n = 2), overdose (n = 2), ruptured abdominal aortic aneurysm (n = 1), pneumonia (n = 1), ischemic bowel (n = 1) and trauma (n = 1). The average acute physiology and chronic health evaluation score II for the ICU patients was 20 ± 8 (range 10 to 33).

Table 1

Table 1

Table 2

Table 2

There was no difference in prebolus PCWP, CO, SV, EDA, and fractional area of contraction values between both groups of patients (Tables 1 and 2). Four patients of 20 responded to fluid administration in the ICU group and 12 patients of 20 in the cardiac surgical group. There were 17 ICU patients receiving inotropic support and only 2 in the cardiac surgical group.

In all patients, the PCWP rose significantly in both responders (12.3 ± 2.2 to 15.4 ± 3.1 mm Hg, P = 0.023) and nonresponders (15.9 ± 3.1 to 21.1 ± 4.2 mm Hg, P < 0.001) (Figure 1, A and B). The prebolus PCWP was significantly smaller in responders when compared with nonresponders (P = 0.003). There was a significant increase in responder EDA after volume administration (15.3 ± 5.4 to 20.1 ± 5.0 cm2, P = 0.026), which was not observed in nonresponders (20.2 ± 4.8 to 21.5 ± 5.3 cm2, P = not significant) (Figure 1, C and D). The prebolus EDA was significantly smaller in responders when compared with nonresponders (P = 0.012).

Figure 1

Figure 1

There was a modest relationship between the SV and EDA in all patients (r = 0.60) (Figure 2A), whereas none was found for SV and PCWP (r = 0.15, Figure 2B).

Figure 2

Figure 2

There was no relationship between EDA and PCWP for cardiac surgical patients (r = 0.35, Figure 3A) and for ICU patients (r = 0.21, Figure 3B). In the cardiac surgical responders (Figure 3A), the prebolus EDA range was 7 to 23 cm2 (n = 12). In the ICU responders (Figure 3B), the prebolus EDA range was 7 to 12 cm2 (n = 4).

Figure 3

Figure 3

Back to Top | Article Outline


PCWP remains the most commonly used variable to assess preload. The EDA is a rapid and direct measurement of LV dimension. We found no correlation between the EDA and the PCWP, which has previously been demonstrated (7,16). In cardiac surgical patients, the EDA has been shown to be sensitive in detecting changes in blood volume (7–10), even in the presence of wall motion abnormalities (7). In surgical patients undergoing hypervolemic hemodilution (8), volume administration resulted in significant increases in SV and EDA, which increased up to a threshold beyond which no further increases were observed. The PCWP, however, continued to rise. In a study in dogs (17), both the LV EDA and the PCWP increased in response to volume administration. These changes correlated well with changes in CO. Further volume administration did not result in any further increases in EDA or CO; however, the PCWP continued to rise. These studies suggest that a threshold or endpoint for fluid administration can be determined for EDA.

We defined responders as those who increased SV by 20% or more. After volume administration, the PCWP increased significantly in responders and nonresponders alike (Figure 1, A and B). There was no significant increase, however, in nonresponder EDA after volume administration. The nonresponders may have already reached their optimal EDA where none or only modest increases in size would be observed. Indeed, nonresponders had a lower LV compliance, as the EDA changed little while the PCWP increased. Some nonresponders, however, did demonstrate an increase in EDA after volume administration. This may have been the result of technical errors in measurement. Alternatively, some patients with poor LV function may have dilated without appreciable increases in SV. Finally, our definition of 20% for increases in SV may have been too high. The nonresponders who increased EDA may have had SV increases slightly less than 20%.

The optimal EDA was difficult to determine as loading conditions, ventricular function, and previous heart disease were not controlled and may have played a large role in determining optimal size. Although responders had significantly smaller prebolus EDA (15.3 ± 5.4 cm2) when compared with nonresponders (20.2 ± 4.8 cm2), the range in both was wide and overlapped significantly (Figure 1, C and D). Furthermore, in patients with wall motion abnormalities or dilated cardiomyopathies, the LV EDA may not have accurately reflected end diastolic volume.

Curiously, a few patients demonstrated a decrease in EDA after volume administration. This may have been the result of technical errors such as probe movement during the study or poor border definition. Alternatively, improvements in loading conditions in response to volume may have resulted in a decreased EDA with little change in SV. The ventilatory cycle may have also influenced the EDA, especially in relatively hypovolemic patients.

We found no correlation between SV and PCWP (Figure 2B). We found only a very modest correlation between the SV and EDA (r = 0.60, Figure 2A) in all our patients. However, in a study examining 16 hemodynamically unstable (mean arterial pressure < 60 mm Hg or cardiac index < 3.0 L · min−1 · /m−2) postoperative patients admitted to an ICU, a good correlation (r = 0.89) was found between the SV and the LV EDA (18). Our poorer relationship may have been the result of a larger proportion of patients with chronic heart disease, who demonstrated large variations in optimal LV sizes and in whom EDA did not reflect true preload. Furthermore, studying a hemodynamically unstable population likely improved the relationship between preload and performance as a larger proportion of patients might have been on the steeper portion of the Starling curve.

It is interesting to note that few ICU patients responded to volume administration when compared with the postoperative cardiac surgical patients (Figure 3). Indeed, it has previously been reported that critically ill patients respond poorly to fluid and show no appreciable increases in LV EDA after volume administration (13). The ICU responders seemed to cluster at an EDA ≤ 12 cm2 (Figure 3B). The small number of responders in the ICU patients could be the result of a septic syndrome, for example, in some patients a high CO would preclude any significant SV increases. Alternatively, overall poor LV performance may have prevented any appreciable recruitment of SV for various LV sizes. Indeed, most of the ICU patients were receiving inotropes when compared with the cardiac surgical group. In the cardiac surgical group, responders were distributed over a wider range of EDA (Figure 3). It was not possible to establish a threshold below which a large proportion of patients respond to volume administration.

In summary, this study provides evidence that the TEE gives information additive to the pulmonary artery catheter in the assessment of preload in an ICU population. Although there may not be a specific threshold EDA value that reliably predicts a response to fluid administration in all patients, the LV EDA may be useful in identifying some critically ill patients who could benefit from volume administration. The decision to administer fluid still relies on clinical judgment.

Back to Top | Article Outline


1. Kaul S, Stratienko AA, Pollock SG, et al. Value of two-dimensional echocardiography for determining the basis of hemodynamic compromise in critically ill patients: a prospective study. J Am Soc Echocardiogr 1994; 7:598–606.
2. Greim CA, Roewer N, Laux G, et al. Transesophageal echocardiography for determining left-ventricular end-diastolic myocardial wall tension. Anasthesiol Intensivved Notfallmed Schmerzther 1994; 29:475–80.
3. Pearson AC, Castello R, Labovitz AJ. Safety and utility of transesophageal echocardiography in the critically ill patient. Am Heart J 1990; 119:1083–9.
4. Porembka D, Hoit B. Transesophageal echocardiography in the intensive care unit. Crit Care Med 1991; 19:826–35.
5. Reichert CLA, Visser CA, Koolen JJ, et al. Transesophageal echocardiography in hypotensive patients after cardiac operations. J Thorac Cardiovasc Surg 1992; 104:321–6.
6. Font VE, Obarski TP, Klein AI. Transesophageal echocardiography in the critical care unit. Cleve Clin J Med 1991; 58:315–22.
7. Cheung AT, Savino JS, Weiss SJ, et al. Echocardiographic and hemodynamic indexes of left ventricular preload in patients with normal and abnormal ventricular function. Anesthesiology 1994; 81:376–87.
8. van Daele MERM, Trouwborst A, van Woerkens LCSM, et al. Transesophageal echocardiographic monitoring of preoperative acute hypervolemic hemodilution. Anesthesiology 1994; 81:602–9.
9. Coriat P, Vrillon M, Perel A, et al. A comparison of systolic blood pressure variations and echocardiographic estimates of end-diastolic left ventricular size in patients after aortic surgery. Anesth Analg 1994; 78:46–53.
10. Reich DL, Konstadt SN, Nejat M, et al. Intraoperative transesophageal echocardiography for the detection of cardiac preload changes induced by transfusion and phlebotomy in pediatric patients. Anesthesiology 1993; 79:10–5.
11. Leung JM, Levine EH. Left ventricular end-systolic cavity obliteration as an estimate of intraoperative hypovolemia. Anesthesiology 1994; 81:1102–9.
12. Wolrab C, Weber T, Tschernich H, et al. Assessment of left ventricular preload: transesophageal echocardiography versus filling pressure. Acta Anaesthesiol Scand Suppl 1997; 111:283–6.
13. Axler O, Tousignant C, Thompson CR, et al. Small hemodynamic effect of typical rapid volume infusions in critically ill patients. Crit Care Med 1997; 25:965–70.
14. Vedrinne JM, Duperret S, Decaillot F, et al. Haemodynamic changes induced by two I: E ratios: a transesophageal echocardiographic study. Can J Anaesth 1997; 44:354–9.
15. Feltes TF, Pignatelli R, Kleinfert S, Mariscalco M. Quantitated left ventricular systolic mechanics in children with septic shock utilizing noninvasive wall-stress analysis. Crit Care Med 1994; 22:1647–58.
16. Thys DM, Hillel Z, Goldman ME, et al. A comparison of hemodynamic indices derived by invasive monitoring and two-dimensional echocardiography. Anesthesiology 1987; 67:630–4.
17. Swenson JD, Harkin C, Pace NL, et al. Transesophageal echocardiography: an objective tool in defining maximum ventricular response to intravenous fluid therapy. Anesth Analg 1996; 83:1149–53.
18. Greim CA, Roewer N, Apfel C, et al. Relation of echocardiographic preload indices to stroke volume in critically ill patients with normal and low cardiac index. Intensive Care Med 1997; 23:411–6.
© 2000 International Anesthesia Research Society