Secondary Logo

Journal Logo

Invited Commentary

Swans in Swine: Cardiac Output Monitoring in Pigs Supported on Venovenous Extracorporeal Membrane Oxygenation

Baronos, Stamatis; Adkins, Kandis

Author Information
doi: 10.1097/MAT.0000000000001763
  • Free

We have read, with great interest, the article by Russ et al.1 regarding their findings that thermodilution based measurements can overestimate cardiac output during veno-venous extracorporeal membrane oxygenation (ECMO) and that the degree of overestimation correlates with recirculation fraction and the ECMO blood flow (QEC)/aortic blood flow (QAO). Conducting this study in actual lung-injured pigs keeps true to significant confounding factors in patients with acute respiratory distress syndrome. It is often supposed that cardiac output monitoring in patients supported on veno-venous ECMO is inaccurate or unreliable, but this well-designed study adds evidence and provides eloquent and thoughtful reasoning for the witnessed observations that support this hypothesis. After controlling for multiple large and small variables, establishing robust reliable monitoring, and performing thoughtful experiments, the authors present a compelling argument on the validity of thermodilution to calculate cardiac output in patients supported on femoral-external jugular ECMO.

The authors are careful to expose potential sources of error or bias, and we must start with the most fundamental, defining the true cardiac output. The authors used the transit-time flow measurement (TTFM) method to measure the true cardiac output. Transit-time flow measurement is a relatively new quantitative volume-flow Doppler technique. This method uses two piezoelectric crystals transmitting ultrasound waves to a reflector on the other side of a blood vessel. The difference between the ultrasound wave transit times between the two crystals is used to calculate blood flow.2 Interestingly, the same size vascular flow probe was used in all subjects for calculation of cardiac output in the ascending aorta despite a noticeable body weight variability (mean body weight 78 ± 9 kg). This may have resulted in differences in aortic size, contour, shape and thickness, not necessarily accounted for and potentially influencing the measured flow.

While an ultrasonographic measurement of ascending aortic flow is monitored here, as the authors note, coronary blood flow separates before the ascending aorta and could represent 50 to 120 ml/min/100 g of myocardial mass. With the average pig heart weighing approximately 300 g, the average heart could divert as much as 360 ml/min to the coronary circulation never reaching the aortic flow sensor. As an alternative, the true cardiac output can be calculated by measuring the left ventricular outflow tract cross sectional area and velocity-time integral using echocardiography3 or as originally described by the dye dilution method.4

The authors found that at higher ECMO flows particularly when QEC/QAO >75%, the thermodilution method overestimates the cardiac output at a greater degree which correlates with the degree of recirculation. Although these findings add to the existing ECMO evidence on recirculation, the absolute values are not generalizable since the exact position of the drainage and return cannula in relation to the right atrium and injection port of the pulmonary artery catheter were not measured or reported. One hypothesis of why the pulmonary artery catheter derived cardiac output (QPa) might overestimate the true cardiac output (in this case the aortic blood flow monitored by direct ultrasonography at the level of the ascending aorta) is the position of the cannulae in proximity to the pulmonary artery catheter. If the position of the drainage or return cannula is close to the injection port, one might guess the QPa may be erroneously high. If the drainage cannula is close to the injection port, part of the injectate can be suctioned in the drainage cannula which can result in a smaller area under the thermodilution curve and thus higher cardiac output. As far as the return cannula is concerned, this blood is exposed to room temperature but is generally externally warmed, and as QEC increases, especially above QAO, the dilution of this bolus at the thermistor will falsely elevate calculated cardiac output.5 Interestingly, the cardiac output was overestimated with thermodilution even when the QEC/QAO was nearly 0%. This may have occurred due to lack of validation of the QAO method. As an alternative, but also using the indicator-dilution technique, relying on dye concentration change as opposed to temperature change may result in more reliable detection of QPa.

The presence of a left-to-right shunt is also not excluded, although could be ruled out by an agitated saline study by echocardiography. This could explain the high QPa/QAO ratio but the incidence of clinically significant left-to-right shunt in pigs is unknown. Lastly, positioning of the pulmonary artery catheter in relation to animal orientation is important. Pigs were placed in the right lateral decubitus position and may have variable lung perfusion, especially in lung-injured pigs who may have asymmetric lung damage and perfusion.

Overall, Russ et al.1 has provided a very well-conducted study seeking to quantify the inaccuracies of thermodilution based measurements of cardiac output, adding to the original body of evidence. A more comprehensive, and ideally minimally invasive, method is needed to capture true cardiac output. Further research is also needed to investigate the effect of thermodilution in other commonly-used cannulation strategies, for example dual-lumen ECMO cannulation, femoral-femoral cannulation, as well as combined dual-lumen veno-venous ECMO with right ventricular support cannulation such as the Protek Duo system (TandemLife, Pittsburg, PA). For now, ECMO care teams have evidence that thermodilution is unreliable in this patient population.


1. Russ M, Steiner E, Boemke W, et al.: Extracorporeal membrane oxygenation blood flow and blood recirculation compromise thermodilution-based measurements of cardiac output. ASAIO J 68: 721–729, 2022
2. Beldi G, Bosshard A, Hess OM, Althaus U, Walpoth BH: Transit time flow measurement: experimental validation and comparison of three different systems. Ann Thorac Surg 70: 212–217, 2000.
3. Huenges K, Pokorny S, Berndt R, Cremer J, Lutter G: Transesophageal Echocardiography in Swine: Establishment of a Baseline. Ultrasound Med Biol 43: 974–980, 2017.
4. Haller M, Zöllner C, Manert W, et al.: Thermodilution cardiac output may be incorrect in patients on venovenous extracorporeal lung assist. Am J Respir Crit Care Med 152(6 Pt 1): 1812–1817, 1995.
5. Reuter DA, Huang C, Edrich T, Shernan S, Eltzschig HK. Cardiac output monitoring using indicator-dilution techniques: basics, limits, and perspectives, Anesth Analg 110:799–811, 2010.

extracorporeal membrane oxygenation; recirculation fraction; cardiac output measurements; thermodilution

Copyright © ASAIO 2022