This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Culp, W. C. Articles by Burnett, C. J. Search for Related Content PubMed PubMed Citation Articles by Culp, W. C., Jr Articles by Burnett, C. J. Related Collections Cardiovascular Monitoring (Cardiac) Technology

---

CARDIOVASCULAR ANESTHESIOLOGY

Validation and Feasibility of Intraoperative Three-Dimensional Transesophageal Echocardiographic Cardiac Output

From the Division of Cardiothoracic Anesthesiology, Department of Anesthesiology, Scott & White Hospital, The Texas A&M University System Health Science Center College of Medicine, Temple, TX.

Address correspondence and reprint requests to W. C. Culp, Jr., MD, Scott & White Hospital, The Texas A&M University System Health Science Center College of Medicine, Department of Anesthesiology, 2401 South 31st St., Temple, TX 76508. Address e-mail to wculp{at}swmail.sw.org.

Abstract

BACKGROUND: In this pilot study, we attempted to validate three-dimensional transesophageal echocardiography (3DTEE) cardiac output and assess its feasibility intraoperatively.

METHODS: Twenty patients undergoing cardiac surgery underwent simultaneous cardiac output determinations during the clinically stable prebypass period by 3DTEE and thermodilution.

RESULTS: The correlation coefficient between cardiac output measured by the two methods was 0.86. The 3DTEE mean bias was 0.27 L/min, limits of agreement –1.64 to 2.17 L/min (approximately ±35%). Three-dimensional data acquisition averaged 43 s; postprocessing took 7 min.

CONCLUSIONS: Three-dimensional TEE can measure cardiac output and is feasible perioperatively. Measurements have good correlation with thermodilution, though with a significant bias and wide limits of agreement.

Cardiac output is a key hemodynamic variable that guides pharmacologic and fluid therapy and is nearly universally measured by thermodilution technique with a pulmonary artery catheter (1), the role of which has been questioned, prompting a search for a minimally invasive means to measure cardiac output (2).

Two-dimensional (2D) echocardiographic techniques can measure stroke volume, but are limited by empiric formulas, which presume left ventricular morphology, and are less accurate than three-dimensional (3D) volumetry (3,4). Doppler-based cardiac output has often, but not always, been shown to provide accurate cardiac output values. Doppler beam misalignment and other factors may lead to velocity under estimation (5–7). Esophageal Doppler probe monitoring may also be used perioperatively, though data on its accuracy are conflicting (8,9). Three-dimensional echocardiographic imaging allows for highly accurate quantification of left ventricular volumes, though typically with long acquisition and postprocessing times, which have hindered its implementation in routine clinical practice (10). Recently, 3D transesophageal echocardiography (TEE) has become more refined and clinically applicable as packaged in a single platform, integrating both traditional ultrasound modalities and 3D imaging software. Advances in volume reconstruction methods have dramatically reduced off-line processing time, and image acquisition has been streamlined (11). By using the minimally invasive, low risk procedure of TEE (12), ventricular volumes can now be measured and cardiac output then derived in a timely fashion. In this pilot study, we sought to validate 3DTEE-derived cardiac output with pulmonary artery catheter thermodilution cardiac output. In addition, we assessed the feasibility of this technique in the busy operating room environment by measuring acquisition and reconstruction times.

METHODS

After receiving local IRB approval and written informed consent, 20 consecutive patients undergoing elective coronary artery bypass grafting and/or aortic valve replacement were chosen for this prospective study. Exclusion criteria included contraindications to TEE or pulmonary artery catheterization, rhythm other than sinus, significant tricuspid regurgitation, or intracardiac shunt. After induction of general anesthesia and before cardiopulmonary bypass, each patient underwent simultaneous cardiac output measurement by 3DTEE and thermodilution. To maintain internal consistency, the same investigator performed 3D volumetry for each patient; likewise, one investigator performed thermodilution for each patient. Each was blinded to the other's results.

Three-dimensional TEE was performed using a standard multiplane probe (V5M, Siemens Medical, Malvern, PA) and a commercially available TEE unit (Acuson CV70®, Siemens Medical) with integrated 3D imaging software (fourSight®, Siemens Medical). Care was taken to optimize and center the left ventricular endocardial border in the mid-esophageal four-chamber view. A series of 2D images was then acquired by increasing the probe's multiplane angle in 5° increments. This automatic process is gated to the electrocardiogram and allows one cardiac cycle to be reconstructed in 3D.

After acquisition, end-diastolic and end-systolic left ventricular volumes were reconstructed. First, end-diastole and end-systole were identified in a series of approximately 15–20 images that comprised the 2D cine loop of one cardiac cycle. Then, the endocardial border was manually traced on four orthogonal plane images for both end-diastole and end-systole (Fig. 1). These four views were automatically selected by the software. A previously validated four-plane model was chosen in an effort to optimize accuracy while minimizing postprocessing time (4). Ventricular shapes were automatically constructed in 3D by applying a spline algorithm to create a closed contour defined by the traced endocardial border. Volumes were calculated by using a "Rotaplane" algorithm, previously described (10). Stroke volume and cardiac output were then derived (stroke volume = end-diastolic volume – end-systolic volume; cardiac output = heart rate x stroke volume). Subjective image quality was graded based on identification of all of the endocardial border in each plane (good), at least 80% of the endocardial border in each plane (fair), or <80% of the endocardial border in each plane (poor).

View larger version (44K): [in this window] [in a new window]   Figure 1. In the multiplanar reformatting screen, each quadrant contains a unique view of the data set at end-systole. The top screens demonstrate two orthogonal views of the left ventricle in long axis. The endocardial border has been manually traced (blue lines). The lower left screen shows the ventricle in short axis. The lower right screen displays the three-dimensional (3D) ventricular cavity and volume as constructed from the traced endocardial borders. This process is repeated for end-diastole, and then the 3D constructs are superimposed.

All patients were monitored with a pulmonary artery catheter (Swan-Ganz Catheter, Edwards Lifesciences, Irvine, CA) and hemodynamic calculations performed and displayed by a computer and monitor system (Siemens SC 7000, Siemens Medical). All patients were placed on a fraction of inspired oxygen (Fio₂) of 1.0 several minutes before measuring cardiac output, and all measurements were made during apnea. Thermodilution using 10 mL of room temperature D5W was begun as 3DTEE acquisition was initiated, and continued until at least three measurements were within 0.75 L/min of each other. The time to obtain at least three thermodilution cardiac outputs was also measured.

For statistical comparison, thermodilution cardiac output was defined as the mean of the three recorded thermodilution values within the preset range of 0.75 L/min. Analysis comparing the cardiac output obtained by the two techniques used the methods proposed by Bland and Altman, paired t-test, and Pearson's correlation coefficient (r) (13).

RESULTS

Patient characteristics and summary data are listed in Table 1. All "good quality" image data were acquired in patients 9–20, suggesting a learning curve for image acquisition. Four acquisitions were not centered adequately for initial volume reconstruction but were subsequently optimized. Reconstructions were performed before comparing with thermodilution results. No complications were noted.

In 15 of the 20 measurements (75%), the two methods gave results within ±1 L/min, and in all cases the two measurements were within ±2 L/min of each other. A scatterplot demonstrates the correlation of the techniques, r = 0.86 (Fig. 2). The difference between the measurements versus the average of the two measurements ("Bland-Altman plot") is shown (Fig. 3). There was a mean bias of 0.27 L/min, and a 95% confidence interval for this mean bias is from –0.19 L/min to 0.72 L/min. Limits of agreement between the two methods is from –1.64 to 2.17 L/min. Though many patients had regional wall motion abnormalities, these patients were not analyzed separately.

View larger version (17K): [in this window] [in a new window]   Figure 2. Scatterplot of three-dimensional transesophageal echocardiography versus thermodilution cardiac output. The middle diagonal line is the line of equality. Values within the outer lines are within 1 L/min of each other, and values outside the outer lines are more than 1 L/min from each other. Pearson's correlation coefficient (r), regression equation, and standard error of the estimate (SEE) are listed.

View larger version (13K): [in this window] [in a new window]   Figure 3. Bland-Altman scatterplot of differences in cardiac output between methods versus average cardiac output of the two methods. The middle horizontal dotted line shows the bias between the methods, which is the mean of the differences. The upper and lower horizontal dotted lines indicate the limits of agreement between the two methods.

DISCUSSION

Cardiac output is an important guide for perioperative fluid and inotropic therapy. Previous work has validated the use of various 3D echocardiographic systems in left ventricular volumetry (10,11); however, none of these studies has assessed intraoperative 3DTEE and its feasibility. Despite high accuracy in nonoperating room environments, intraoperative 3D volumetry has not yet become commonplace because of cumbersome acquisition, long postprocessing times, and the suboptimal integration of the 3D software and the ultrasound machine. Using the transesophageal approach in our study, acquisition was 43 s with postprocessing times of approximately 7 min using manual border detection. This is a marked advance in comparison to the acquisition and postprocessing times of 4 and 17 min, respectively, as described in Van den Bosch et al.'s study using a "real-time" transthoracic approach (10). Nevertheless, in the patient with a Swan-Ganz catheter already in place, thermodilution is clearly the faster of the two techniques, and is less likely to distract the anesthesiologist from patient care. However, automated border detection algorithms already developed for transthoracic volumetry (14) may soon be applied to the transesophageal approach, potentially reducing postprocessing time and sonographer workload, but with an unknown effect on accuracy. Furthermore, the technological leap of real-time 3DTEE holds the promise of much faster data acquisition, potentially offering stroke volume values within seconds. There is a prototype real-time 3DTEE probe (15), and it is expected to be commercially available soon.

Transthoracic and transesophageal 3D volumetry are inherently limited by temporal resolution in tachycardic patients, potentially underestimating stroke volume (16). This occurs because the series of 10–20 2D images that comprise one cardiac cycle are less likely to capture the peak end-diastolic volume as the heart rate increases. This leads to a falsely low volume measurement. Similarly, end-systolic volume may be overestimated if not precisely timed. This can cause a systematic underestimation of stroke volume, and therefore, cardiac output. Many 3D echocardiographic studies underestimate stroke volume, and this may be due to limited temporal resolution. Future studies should determine heart rate thresholds of accuracy. Sinus rhythm is also necessary for accurate electrocardiogram gating using current technology, but may become less important as real-time 3DTEE is developed. Our real-world operating room environment data suggest a strong correlation with thermodilution. However, the overall accuracy of the technique with the current ultrasound system may not be adequate for most clinical situations, and has not been validated during periods of dynamic hemodynamic changes.

Other limitations of this technique include the lack of applicability for prolonged postoperative monitoring, the need for specialized equipment and personnel, and the requirement of accurate endocardial border imaging. Good quality endocardial border imaging was achieved in 7 of the last 11 studied patients (64%), whereas no good quality images were obtained in the first 9 patients, reflecting what we suspect is the learning curve for the technique. This learning curve effect was detected only after post hoc analysis and is a limitation of our study. Absolute optimization of the 2D ventricular image is key to high quality 3D acquisition. Use of echogenic contrast agents in this application to further enhance imaging may lead to improved accuracy, and future research in this area is indicated (17). The use of additional plane pairs may also slightly increase accuracy (4), though at a large expense of postprocessing time; doubling our four-plane model to eight planes led to an average increase in postprocessing time of 4 min (total time 12 min), and doubling again to 16 planes took an additional 6–7 min (total time 18–19 min).

One echocardiographic alternative to measuring cardiac output is a Doppler-based method measuring blood flow through the aortic valve. This technique takes approximately half the time of 3D volumetry described in our study, and has, in general, good accuracy as compared with thermodilution. Perrino et al. report excellent correlation (r = 0.98, limits of agreement ±1.1 L/min), although others suggest that neither 2D nor Doppler TEE techniques are adequately reliable, with a Doppler bias of –0.36 to 0.5 L/min and a percentage error of 37.0%–42.5% (5–7). Our pilot study did not compare Doppler techniques with 3D volumetry. However, larger future studies would be enhanced by comparing all three techniques in the pre- and postbypass periods as 3D volumetry is refined.

In conclusion, 3DTEE cardiac output is a feasible method of assessing heart function in stable cardiac surgery patients before bypass, though in its current form it remains more time consuming and less accurate than Doppler techniques and its applicability in less stable situations has yet to be investigated. It correlates fairly well with the clinical "gold standard" of thermodilution, though with a mean bias of 0.27 L/min and limits of agreement of ±1.9 L/min (about 35%). Evolutions in automated processing and real-time 3DTEE are anticipated to advance the intraoperative utility of this technique. With these refinements, 3DTEE volumetry may play a significant role in the minimally invasive assessment of perioperative hemodynamics.

ACKNOWLEDGMENTS

The authors thank Mark W. Riggs, Ph.D. (Associate Professor, Department of Mathematics and Computer Science, Abilene Christian University, Abilene, TX, USA) for his statistical analysis, and David P. Ciceri, M.D. (Assistant Professor, Department of Anesthesiology, The Texas A&M University System Health Science Center College of Medicine, Scott & White Hospital, Temple, TX, USA) for his help in study design.

Footnotes

Accepted for publication July 19, 2007.

The authors have no conflicts of interest to declare.

REFERENCES

1. Espersen K, Jensen W, Rosenborg D, Thomsen JK, Eliasen K, Olsen NV, Kanstrup IL. Comparison of cardiac output measurement techniques: thermodilution, Doppler, CO2-rebreathing and the direct Fick method. Acta Anaesthesiol Scand 1995;39:245–51[Web of Science][Medline]
2. Dalen JE, Bone RC. Is it time to pull the pulmonary artery catheter? JAMA 1996;276:916–18[Abstract/Free Full Text]
3. Lang RM, Bierig M, Devereux RB, Flachskampf FA, Foster E, Pellikka PA, Picard MH, Roman MJ, Seward J, Shanewise JS, Solomon SD, Spencer KT, Sutton MS, Stewart WJ. Recommendations for chamber quantification: a report from the American Society of Echocardiography's Guidelines and Standards Committee and the Chamber Quantification Group, developed in conjunction with the European Association of Echocardiography, a branch of the European Society of Cardiology. J Am Soc Echocardiogr 2005;18:1440–63[Web of Science][Medline]
4. Gutiérrez-Chico JL, Zamorano JL, Pérez de Isla L, Orejas M, Almería C, Rodrigo JL, Ferreirós J, Serra V, Macaya C. Comparison of left ventricular volumes and ejection fractions measured by three-dimensional echocardiography and cardiac magnetic resonance in patients with various cardiomyopathies. Am J Cardiol 2005;95:809–13[Web of Science][Medline]
5. Perrino AC Jr, Harris SN, Luther MA. Intraoperative determination of cardiac output using multiplane transesophageal echocardiography: a comparison to thermodilution. Anesthesiology 1998;89:350–7[Web of Science][Medline]
6. Bettex DA, Hinselmann V, Hellerman JP, Jenni R, Schmid ER. Transoesophageal echocardiography is unreliable for cardiac output assessment after cardiac surgery compared with thermodilution. Anaesthesia 2004;59:1184–92[Web of Science][Medline]
7. Coats AJ. Doppler ultrasonic measurement of cardiac output: reproducibility and validation. Eur Heart J 1900;11:539–50
8. Lafanechére A, Albaladejo P, Raux M, Geeraerts T, Bocquet R, Wernet A, Castier Y, Marty J. Cardiac output measurements during infrarenal aortic surgery: echo-esophageal Doppler versus thermodilution catheter. J Cardiothorac Vasc Anesth 2006;20:26–30[Web of Science][Medline]
9. Sharma J, Bhise M, Singh A, Mehta Y, Trehan N. Hemodynamic measurements after cardiac surgery: transesophageal Doppler versus pulmonary artery catheter. J Cardiothorac Vasc Anesth 2005;19:746–50[Web of Science][Medline]
10. Van den Bosch AE, Robbers-Visser D, Krenning BJ, Voormolen MM, McGhie JS, Helbing WA, Roos-Hesselink JW, Simoons ML, Meijboom FJ. Real-time transthoracic three-dimensional echocardiographic assessment of left ventricular volume and ejection fraction in congenital heart disease. J Am Soc Echocardiogr 2006;19:1–6[Medline]
11. Kühl HP, Franke A, Merx M, Hoffman R, Puschmann D, Hanrath P. Rapid quantification of left ventricular function and mass using transesophageal three-dimensional echocardiography: validation of a method that uses long-axis cutplanes. Eur J Echocardiogr 2000;1:213–21[Medline]
12. Min JK, Spencer KT, Furlong KT, DeCara JM, Sugeng L, Ward RP, Lang RM. Clinical features of complications from transesophageal echocardiography: a single-center case series of 10,000 consecutive examinations. J Am Soc Echocardiogr 2005;18:925–9[Web of Science][Medline]
13. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10[Web of Science][Medline]
14. Kühl HP, Schreckenberg M, Rulands D, Katoh M, Schäfer W, Schummers G, Bücker A, Hanrath P, Franke A. High-resolution transthoracic real-time three-dimensional echocardiography: quantitation of cardiac volumes and function using semi-automatic border detection and comparison with cardiac magnetic resonance imaging. J Am Coll Cardiol 2004;43:2083–90[Abstract/Free Full Text]
15. Pua EC, Idriss SF, Wolf PD, Smith SW. Real-time 3D transesophageal echocardiography. Ultrason Imaging 2004;26:217–32[Web of Science][Medline]
16. Müller S, Bartel T, Baumann G, Erbel R. Preliminary report: evaluation of three-dimensional echocardiographic volumetry by simultaneous thermal dilution in coronary artery disease. Cardiology 1996;87:552–9[Web of Science][Medline]
17. Handke M, Heinrichs G, Magosaki E, Lutter G, Bode C, Geibel A. Three-dimensional echocardiographic determination of cardiac output at rest and under dobutamine stress: comparison with thermodilution measurements in the ischemic pig model. Echocardiography 2003;20:47–55[Web of Science][Medline]

This article has been cited by other articles:

				F. Mahmood, A. Christie, and R. Matyal Transesophageal Echocardiography and Noncardiac Surgery Seminars in Cardiothoracic and Vascular Anesthesia, December 1, 2008; 12(4): 265 - 289. [Abstract] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Culp, W. C. Articles by Burnett, C. J. Search for Related Content PubMed PubMed Citation Articles by Culp, W. C., Jr Articles by Burnett, C. J. Related Collections Cardiovascular Monitoring (Cardiac) Technology