Ann Thorac Surg 2001;71:868-871
© 2001 The Society of Thoracic Surgeons
Original article: cardiovascular
Left ventricular dysfunction during extracorporeal membrane oxygenation in a hypoxemic swine model
Irving Shen, MDa,
Fiona H. Levy, MDb,
Arnold M. Benak, CCPa,
Christine L. Rothniea,
P. Pearl O’Rourke, MDb,
Brian W. Duncan, MDa,
Edward D. Verrier, MDa
a Division of Cardiothoracic Surgery, University of Washington, Seattle, Washington, USA
b Department of Anesthesia, Children’s Hospital and Medical Center, Seattle, Washington, USA
Accepted for publication June 8, 2000.
Address reprint requests to Dr Verrier, Division of Cardiothoracic Surgery, Department of Surgery, University of Washington, Box 356310, 1959 NE Pacific St, Seattle, WA 98195-6310
e-mail: verrier{at}ctd.surgery.washington.edu
 |
Abstract
|
|---|
Background. Perfusion of the coronary circulation with hypoxemic blood from the left ventricle has been postulated to cause myocardial dysfunction during venoarterial extracorporeal membrane oxygenation for respiratory support.
Methods. We investigated this hypothesis in 10 anesthetized open-chest piglets (7 to 9 kg) undergoing venoarterial extracorporeal membrane oxygenation after placement of minor-axis sonomicrometry crystals and left ventricular micromanometer. The left atrial partial pressure of oxygen was made hypoxemic (25 to 40 mm Hg) after initiation of extracorporeal membrane oxygenation by ventilation with a hypoxic gas mixture. Left ventricular contractile function, including peak LV pressure, shortening fraction, maximum rate of increase of left ventricular pressure, velocity of circumferential fiber shortening, end-systolic pressure–minor axis dimension relationship, and preload recruitable dimensional stroke work, was measured or calculated on extracorporeal membrane oxygenation before (baseline) and at 4 and 6 hours after rendering the left atrial blood hypoxemic.
Results. Left ventricular shortening fraction and velocity of circumferential fiber shortening were significantly lower (p < 0.05) at 4 and 6 hours when compared with baseline. The slope of the end-systolic pressure–minor axis dimension relationship decreased but was not significantly different at 4 and 6 hours when compared with baseline owing to poor linear correlation (r = 0.30 to 0.93). The preload recruitable dimensional stroke work was more linear (r = 0.87 to 0.99), and the slope was significantly lower (p < 0.01) at 4 and 6 hours when compared with baseline.
Conclusions. Hypoxemic cardiac output from the left ventricle during venoarterial extracorporeal membrane oxygenation is associated with depression of left ventricular systolic function in this animal model. Current use of venoarterial extracorporeal membrane oxygenation for respiratory support may not provide adequate oxygen supply to the myocardium.
|
Introduction
|
|---|
Venoarterial extracorporeal membrane oxygenation (VA ECMO) is an established mode of cardiopulmonary support for neonates with severe respiratory failure [1, 2]. While undergoing VA ECMO for respiratory failure, these patients can often exhibit varying degrees of cardiac dysfunction [3–5], which has been referred to as the "cardiac stun syndrome." Clinical echocardiographic studies have documented changes in left ventricular (LV) systolic performance during ECMO, including decreases in LV shortening fraction, pulmonary arterial and aortic flow velocities, stroke volume, and velocity of circumferential fiber shortening [6, 7]. These changes can persist for days even in the presence of a normalized systemic arterial oxygen tension (PO2) [3]. Some factors, including reoxygenation injury [8] and inflammatory mediators [9], have been proposed to be the cause of these changes in the clinical setting but the origin of this phenomenon remains unclear. Understanding the factor(s) that influence cardiac function during ECMO is important because ECMO is now becoming more accepted as a modality for extracorporeal support in pediatric patients that are in cardiac failure after congenital heart surgery [10–13].
Recent studies using radioactive microspheres [14–16] and shunt equation [17] showed that coronary perfusion during ECMO is predominantly by blood from the LV when bypass flow rate is less than total cardiac output [17]. This means that coronary perfusion during VA ECMO in patients with respiratory failure is predominantly with hypoxemic blood returning from the pulmonary circulation. Hence, we postulated that the perfusion of the coronary circulation with desaturated blood from the LV may in part contribute to the decrease in ventricular function that has been reported to occur during VA ECMO.
|
Material and methods
|
|---|
Ten farm piglets of either sex weighing between 7 and 9 kg (mean, 8.7 kg) were used. The experimental preparation, data acquisition, and data analysis have been described in detail previously [18]. During the surgical preparation and initially after placing the animal on ECMO, left atrial (LA) and systemic PO2 were maintained between 100 and 200 mm Hg by adjusting the PO2 in the ventilator and oxygenator. Systemic pH was maintained throughout the experiment by adjusting the minute ventilation of the animal, the rate of sweep gas of the ECMO oxygenator, and administering intravenous sodium bicarbonate. After instrumentation, VA ECMO was initiated, and extracorporeal flow rate was maintained at approximately 50% of a thermodilution-derived pre-ECMO baseline cardiac output. After the animal stabilized on ECMO and baseline data were collected, the LA PO2 was then decreased and maintained at 25 to 40 mm Hg by ventilating the animal with an inspiratory gas mixture of nitrogen and oxygen. The systemic arterial PO2, however, was maintained between 100 and 200 mm Hg throughout the experiment by adjusting the PO2 in the oxygenator and ECMO flow rate.
While on ECMO, data were collected before (baseline) and at 4 and 6 hours after rendering LA blood hypoxemic. Qualitative two-dimensional epicardial echocardiography was performed before and during ECMO to rule out any distortion of LV shape.
This study was reviewed and approved by the Animal Care Committee of our institution. All animals received humane care in compliance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research and the "Guide for Care and Use of Laboratory Animals" (National Institutes of Health, publication 86-23, revised 1985).
|
Results
|
|---|
All 10 animals were successfully placed on VA ECMO, and the average ECMO flow rate was 78.4 ± 3.0 mL · kg-1 · min-1. However, 2 animals experienced severe global biventricular dysfunction that progressed to electromechanical dissociation approximately 4 hours after rendering the LA blood hypoxemic. Metabolic causes of dysfunction were not identified in these 2 animals. The data from these 2 animals were excluded from further analysis and therefore our results reported here are based on data obtained from the remaining 8 animals.
Qualitative two-dimensional epicardial echocardiography performed during data acquisition did not show any evidence of septal flattening or deviation, which can distort the concentricity of the LV and invalidate our function data. The average number of cardiac cycles obtained for analysis during each data acquisition was 13 (range, 6 to 23 cycles). Mean hemodynamics and ejection phase indices in Table 1 showed no significant changes in heart rate and end-diastolic dimension at 4 and 6 hours when compared with baseline. Maximum rate of increase of LV pressure showed a gradual decline at 4 and 6 hours although it did not achieve statistical significance. Peak LV pressure likewise showed a gradual decline that became statistically significant (p < 0.05) at 6 hours after decreasing the LA PO2. Percent LV shortening fraction and velocity of circumferential fiber shortening decreased significantly (p < 0.05) at 4 and 6 hours after reducing LA PO2.
The slope, diameter axis intercept, and correlation coefficient values for the end-systolic pressure–minor axis dimension relationship (ESPDR) are listed in Table 2. Although the slope and x intercept showed a large decrease at 4 and 6 hours, the differences were not statistically significant when compared with baseline. This is most likely because of the poor linear correlation (r = 0.30 to 0.93), which resulted in a wide scattering and overlapping of data.
Table 3 shows the slope, diameter axis intercept, and correlation coefficient values for the preload recruitable dimensional stroke work (PRDSW) relationship. Slope at 4 and 6 hours was significantly decreased when compared with baseline (p < 0.01). Mean diameter axis intercept showed an increase but it was not statistically significant. The PRDSW relationship was much more linear than the ESPDR.
|
Comment
|
|---|
The phenomenon known as cardiac stun syndrome has been documented to occur clinically in patients while on VA ECMO for primary respiratory failure [3, 5]. This syndrome is characterized by severe cardiac dysfunction shortly after being placed on VA ECMO, with the heart exhibiting normal electrical activity but marked decrease in pulse pressure and cardiac contractility. Progression to a severe form of this dysfunction has been known to result in electromechanical dissociation [4] with the total cardiac output being entirely derived from the ECMO circuit alone.
Venoarterial ECMO augments systemic oxygenation in patients with severe respiratory failure, and cardiac performance in these patients should theoretically improve because their heart are structurally normal. However, sustained and even progressive deterioration of cardiac function in these patients after initiation of ECMO can be related to multiple causes including myocardial injury after ischemia and reoxygenation. Recent studies [14–17] have suggested that inadequate delivery of oxygenated blood from the ECMO circuit to the coronary circulation may also play a role in the cardiac stun syndrome. Because VA ECMO is a form of partial cardiopulmonary bypass, a small but significant amount of blood still passes through the pulmonary circulation and returns to the LV. This hypoxic blood is ejected from the LV and may provide a significant contribution to coronary perfusion while on ECMO.
The goal of our experiment was to evaluate whether current uses of VA ECMO for primary respiratory failure contributes to a decrease in cardiac performance. Our model simulates the clinical picture of primary pulmonary oxygenation failure, but the structurally normal heart was not subjected to any ischemic injury before ECMO. This eliminates the possible residual effects of myocardial ischemia and reoxygenation injury. We noticed a similar reduction in ejection phase indices, including LV shortening fraction, peak LV pressure, and velocity of circumferential fiber shortening during ECMO as seen clinically [5–7]. These factors, however, may be less reliable inasmuch as they are influenced by changes in loading condition and heart rate. Furthermore, these changes were also seen in animals placed on VA ECMO alone without rendering LA PO2 hypoxemic during the experiment [18].
We used the ESPDR and PRDSW relationship as load-insensitive indices of cardiac contractility. This allowed us to study the intrinsic inotropic state of the heart in the absence of confounding variables such as preload, afterload, or heart rate. The ESPDR is a modification of the traditional end-systolic pressure-volume relationship in which end-diastolic minor axis diameter is used instead of volume. This linear relationship has been shown to be equally responsive to global changes in LV contractility [19] as reflected by a shift in the slope and x intercept. We demonstrated in our study a decrease in the slope and x intercept of the ESPDR. Neither of these changes, however, was statistically significant because of poor linear correlation. The scattering of data points is probably caused by a shift from a linear to a curvilinear relationship. This change in ESPDR has since been shown in experimental studies to occur in hearts subjected to ischemia or to a decrease in LV end-diastolic volume [20, 21]. This may explain the variability we saw in linear regression coefficients and as such may limit the practical usefulness of this method as a means to evaluate cardiac contractility in this model.
The relationship between cardiac stroke work and end-diastolic minor axis diameter (PRDSW) has been proposed to be a reliable, load-insensitive, and highly linear measure of cardiac contractility even after subjecting the LV to a moderate amount of ischemia and reperfusion injury [22, 23]. Our study showed a significant diminution of intrinsic cardiac contractile function with a significant decrease in the slope of the PRDSW relationship. Although the diameter axis intercept did not show a concomitant increase, the slope of the PRSW relationship has been suggested to be a more reliable and sensitive indicator to changes in contractile function [24]. A decrease in the slope of PRDSW suggested that lowering the LA PO2 did cause a significant decrease in the intrinsic myocardial contractile function that was independent of loading conditions. Furthermore, this decrease in function is unlikely caused by deterioration of the model as similar preparations without LA PO2 hypoxemia did not show a decrease in cardiac function after 6 hours of ECMO [18].
Two animals in our study developed profound ventricular dysfunction resulting in electromechanical dissociation without obvious metabolic or physical causes. This may represent a severe form of hypoxic injury to the myocardium. The ventricle did not respond to direct volume infusion into the LV through the LA catheter. Therefore, electromechanical dissociation in this setting was presumed to be caused by hypoxic injury to the LV as is seen occasionally in patients during clinical VA ECMO.
Even though this study model simulates clinical ECMO for patients with severe respiratory failure, it is limited by its relatively brief duration of support (6 hours) as compared with prolonged duration of ECMO support used clinically. However, measurable impairment of ventricular function was already observed during this brief period of ECMO support, presumably because of hypoxic coronary perfusion. This suggests that a more profound and possibly permanent deterioration of LV function may result from a more prolonged course of ECMO support. A further limitation of this model is that these experimental animals had normal cardiorespiratory function and anatomy before support. This model represents a best-case scenario in that patients with cardiorespiratory failure presupport might be expected to tolerate hypoxia-induced myocardial dysfunction even worse than demonstrated in this model. Certainly myocardial dysfunction secondary to hypoxic coronary perfusion would be poorly tolerated in children with preexisting cardiac disease that require ECMO for circulatory support.
This study has significant implications for clinical management. Children on ECMO often are managed with minimal ventilator support to "rest the lungs" during recovery from respiratory failure. Based on this study, hypoxemia in the pulmonary venous return with native ventricular ejection providing significant coronary perfusion might be a significant cause of progressive myocardial dysfunction. These results suggest that an attempt should be made at maintaining normal pulmonary venous oxygen level when there is significant native ventricular ejection. Attempting to limit ventricular ejection by increasing ECMO flows is an alternative approach that imparts the theoretical risk of significantly increasing afterload to native ventricular ejection. From these results it would appear advisable to provide some ventilator support to maintain LA PO2 greater than 100 mm Hg. These levels would probably be attainable with moderate levels of ventilator support and would remove this as an etiologic factor in myocardial stunning during VA ECMO.
We conclude that oxygen delivery to the myocardium during VA ECMO as used currently for primary respiratory failure may be suboptimal even though indicators of systemic oxygen delivery to other vital organs may be adequate. A substantial portion of coronary perfusion may be provided by desaturated blood from the LV that has circulated through the pulmonary bed and returns to the LV still hypoxemic. This may contribute to the sustained myocardial dysfunction seen clinically in some patients on VA ECMO.
|
References
|
|---|
-
O’Rourke P.P., Crone R.K., Vacanti J.P., et al. Extracorporeal membrane oxygenation and conventional medical therapy in neonates with persistent pulmonary hypertension of the newborn: a prospective randomized study. Pediatrics 1989;84:957-963.[Abstract/Free Full Text]
-
Bartlett R.H., Roloff D.W., Cornell R.G., Andrews A.F., Dillon P.W., Zwischenberger J.B. Extracorporeal circulation in neonatal respiratory failure: a prospective randomized study. Pediatrics 1985;76:479-487.[Abstract/Free Full Text]
-
Hirschl R.B., Heiss K.F., Bartlett R.H. Severe myocardial dysfunction during extracorporeal membrane oxygenation. J Pediatr Surg 1992;27:48-53.[Medline]
-
Rosenberg E.M., Cook L.N. Electromechanical dissociation in newborns treated with extracorporeal membrane oxygenation: an extreme form of cardiac stun syndrome. Crit Care Med 1991;19:780-784.[Medline]
-
Martin G.R., Short B.L., Abbott C., O’Brien A.M. Cardiac stun in infants undergoing extracorporeal membrane oxygenation. J Thorac Cardiovasc Surg 1991;101:607-611.[Abstract]
-
Kimball T.R., Daniels S.R., Weiss R.G., et al. Changes in cardiac function during extracorporeal membrane oxygenation for persistent pulmonary hypertension in the newborn infant. J Pediatr 1991;118:431-436.[Medline]
-
Martin G.R., Short B.L. Doppler echocardiographic evaluation of cardiac performance in infants on prolonged extracorporeal membrane oxygenation. Am J Cardiol 1988;62:929-934.[Medline]
-
Holley D.G., Short B.L., Karr S.S., Martin G.R. Mechanisms of changes in cardiac performance in infants undergoing extracorporeal membrane oxygenation. Crit Care Med 1994;22:1865-1870.[Medline]
-
Kirklin J.K., Westaby S., Blackstone E.H., Kirklin J.W., Chenoweth D.E., Pacifico A.D. Complement and the damaging effects of cardiopulmonary bypass. J Thorac Cardiovasc Surg 1983;86:845-857.[Abstract]
-
Ziomek S., Harrell J.E., Fasules J.W., et al. Extracorporeal membrane oxygenation for cardiac failure after congenital heart operation. Ann Thorac Surg 1992;54:861-867.[Abstract]
-
Rogers A.J., Trento A., Siewers R.D., et al. Extracorporeal membrane oxygenation for postcardiotomy cardiogenic shock in children. Ann Thorac Surg 1989;47:903-906.[Abstract]
-
Raithel S.C., Pennington G., Boegner E., Fiore A., Weber T.R. Extracorporeal membrane oxygenation in children after cardiac surgery. Circulation 1992;86(Suppl 2):II305-II310.
-
Ferrazzi P., Glauber M., Domenico A.D., et al. Assisted circulation for myocardial recovery after repair of congenital heart disease. Eur J Cardiothorac Surg 1991;5:419-424.[Abstract]
-
Nowlen T.T., Salley S.O., Whittlesey G.C., et al. Regional blood flow distribution during extracorporeal membrane oxygenation in rabbits. J Thorac Cardiovasc Surg 1989;98:1138-1143.[Abstract]
-
Kinsella J.P., Gerstmann D.R., Rosenberg A.A. The effect of extracorporeal membrane oxygenation on coronary perfusion and regional blood flow distribution. Pediatr Res 1992;31:80-84.[Medline]
-
Smith H.G., Whittlesey G.C., Kundu S.K., et al. Regional blood flow during extracorporeal membrane oxygenation in lambs. ASAIO Trans 1989;35:657-660.[Medline]
-
Secker-Walker J.S., Edmonds J.F., Spratt E.H., Conn A.W. The source of coronary perfusion during partial bypass for extracorporeal membrane oxygenation (ECMO). Ann Thorac Surg 1976;21:138-143.[Abstract]
-
Shen I., Levy F.H., Vocelka C.R., Akimoto H., Thomas R., O’Rourke P.P., Verrier E.D. The effect of veno-arterial extracorporeal membrane oxygenation on the left ventricular systolic function of swine. J Am Coll Card 1993;21:191A.
-
Little W.C., Freeman G.L., O’Rourke R.A. Simultaneous determination of left ventricular end-systolic pressure-volume and pressure-dimension relationships in closed-chest dogs. Circulation 1985;71:1301-1308.[Abstract/Free Full Text]
-
Toombs C.F., Vinten-Johansen J., Yokoyama H., Johnston W.E., Julian J.S., Cordell A.R. Nonlinearity of indexes of left ventricular performance: effects on estimation of slope and diameter axis intercepts. Am J Physiol 1991;260:H1802-H1809.[Abstract/Free Full Text]
-
Crottogini A.J., Willshaw P., Barra J.G., Armentano R., Cabrera Fischer E.I., Pichel R.H. Inconsistency of the slope and the volume intercept of the end-systolic pressure-volume relationship as individual indexes of inotropic state in conscious dogs: presentation of an index combining both variables. Circulation 1987;76:1115-1126.[Abstract/Free Full Text]
-
Glower D.D., Spratt J.A., Snow N.D., et al. Linearity of the Frank-Starling relationship in the intact heart: the concept of preload recruitable stroke work. Circulation 1985;71:994-1009.[Abstract/Free Full Text]
-
Glower D.D., Spratt J.A., Kabas J.S., Davis J.W., Rankin J.S. Quantification of regional myocardial dysfunction after acute ischemic injury. Am J Physiol 1988;255:H85-H93.[Abstract/Free Full Text]
-
Pagel P.S., Kampine J.P., Schmeling W.T., Warltier D.C. Comparison of end-systolic pressure-length relations and preload recruitable stroke work as indices of myocardial contractility in the conscious and anesthetized, chronically instrumented dog. Anesthesiology 1990;73:278-290.[Medline]
This article has been cited by other articles:
|
|
|
|
 
B. W. Duncan
Mechanical circulatory support for infants and children with cardiac disease
Ann. Thorac. Surg.,
May 1, 2002;
73(5):
1670 - 1677.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
