| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2001;71:862-867
© 2001 The Society of Thoracic Surgeons
a Division of Cardiothoracic Surgery, University of 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 |
|---|
 Top Abstract IntroductionMaterial and methodsResultsCommentReferences |
|---|
Methods. Eight piglets (5 to 11 kg) were anesthetized, instrumented, and placed on VA ECMO. Throughout the experiment, systemic and left atrial partial pressure of oxygen were maintained between 100 to 200 mm Hg. At ECMO flow rate of 50% of baseline cardiac output, data were collected prior to ECMO, at 4 and 6 hours during ECMO, and after weaning from ECMO. Data measured or calculated for each time point included heart rate, LV pressures and minor axis dimensions at different preloads, first derivative of LV pressure with respect to time, velocity of circumferential fiber length shortening (VCF), LV shortening fraction (LVSF), ESPDR, and PRDSW.
Results. A significant (p < 0.05) decrease in LVSF and VCF was seen at 4 and 6 hours during ECMO when compared to baseline, but the ESPDR and PRDSW did not change during ECMO.
Conclusions. VA ECMO alone changes some of the load-dependent parameters of contractility, but intrinsic function of the heart is not significantly affected as measured by load-insensitive indices of LV performance.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
The clinical manifestations of cardiac stunning during ECMO is characterized by a marked decrease in, or loss of, aortic pulse pressure at less than anticipated total bypass flow rate and by an elevation of the patient’s systemic PaO2 approaching that of the oxygenator (350 to 400 mm Hg) [5, 7]. Echocardiographic and pulse Doppler findings may include a dilated left ventricle (LV) with decreased shortening and ejection fractions, diminished pulsatility of aortic blood flow, and decreased LV velocity of circumferential fiber shortening (VCF) [7, 8]. Because these echocardiographic data are strongly influenced by changes in loading conditions, and as the use of VA ECMO can alter both preload or afterload, their validity as indices of intrinsic LV contractility has been questioned [8, 9]. Furthermore, it is unclear whether concomitant clinical conditions like sepsis, acidosis, or hypoxia contribute to the decrease in cardiac performance. Results of clinical studies [10–12] examining the relationship between heart rate–corrected velocity of circumferential fiber shortening (VCFc) and end-systolic wall stress (ESS) as a load-independent measure of cardiac function during the use of VA ECMO were inconclusive.
Very little is known in the laboratory setting about the effects of VA ECMO on contractility of the LV. Bavaria and colleagues [9] demonstrated that VA ECMO decreases LV wall stress in normal hearts. However, a progressive rise in wall stress in postischemic hearts occurred with an increase in ECMO flow rate as a result of a concomitant increase in afterload. This increase in afterload, however, was offset by an improvement in myocardial contractility while on ECMO. Other experimental studies in healthy animals have demonstrated a decrease in LV function during ECMO suggesting that bypass itself may have affected contractility of the heart [13, 14].
The possibility of ECMO inducing cardiac dysfunction is important because ECMO is increasingly used for postoperative circulatory support in pediatric patients after complex congenital cardiac defect repair. The existing data provide a confusing picture on the effect of VA ECMO on intrinsic LV function. We have developed an animal model for VA ECMO that allows us to assess intrinsic LV function using two load-independent measures, preload recruitable dimensional stroke work (PRDSW) and end-systolic pressure dimension relationship (ESPDR), as well as some of the previously reported load-dependent indices such as first derivative of LV pressure with respect to time (dP/dt), LV shortening fraction (LVSF), and VCF.
|
Material and methods |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
After obtaining adequate anesthesia and premedication with gentamicin (2 mg/kg) and cefazolin (50 mg/kg), the heart was exposed through a median sternotomy under sterile conditions. Silastic pneumatic occluders were placed around the superior and inferior venae cavae for changing LV preload during data acquisition. Changes in LV dimension were measured by a pair of minor axis epicardial ultrasonic dimension transducers (HE3-2, Triton Technology, San Diego, CA) coupled to a sonomicrometer (Model 120, Triton Technology). Left ventricular pressure (LVP) was obtained from a high- fidelity 5F micromanometer catheter (Model SPC-450, Millar Instruments, Houston, TX) that was placed into the LV cavity through a stab wound in the apex. A polyethylene catheter was introduced via the left atrial appendage into the left atrium for chamber pressure measurements and left heart blood gas sampling.
After intravenous systemic heparinization (300 units/kg), ECMO venous cannulae were placed in the jugular vein and femoral veins. An arterial cannula was placed in the common carotid artery. The tip of the arterial cannula was positioned in the ascending aorta approximately 1 to 2 centimeters distal to the aortic valve. Activated clotting time (ACT) was measured every hour, and doses of intravenous heparin sodium were given to maintain ACT between 180 to 240 seconds during extracorporeal circulation.
The ECMO circuit was assembled and primed according to the methods described by Bartlett and coworkers [15]. Venous blood was drained by gravity into a reservoir bladder (R30, AVECOR, Plymouth, MN) and was circulated by a precalibrated Stockert/Shiley rollerhead pump through a membrane oxygenator (0800, AVECOR) and a heat exchanger (P-7-14, AVECOR). The experimental circuit is illustrated in Figure 1.
|
Data were collected before (baseline), at 4 and 6 hours during, and after weaning from ECMO (post-ECMO). All animals were allowed to stabilize for at least 30 minutes before each data acquisition. For each data acquisition run, five channels of analog signals, including electrocardiogram, systemic arterial pressure, LV minor axis dimension, LVP, and LV dP/dt, were filtered through a 50 Hz low pass filter, digitized at 200 Hz by a high-speed A-D converter (Model 1012, ADAC, Woburn MA), and stored on a digital disk for later analysis. Two to three sets of data were obtained at each time point under a steady state condition and over a physiologic range of LVP. Data were recorded during transient vena caval occlusion or by administering volume to change LV preload. The animals were held at end expiration during data recording.
Qualitative two-dimensional epicardial echocardiography was performed before and during ECMO to look for distortion of the LV shape which can thereby invalidate the function data. Serum potassium and ionized calcium levels were measured at the time of data acquisition and data were excluded if any of these measurements were outside of the normal swine physiologic ranges of 3.7 to 5.2 mEq/L and 2.32 to 2.58 mmol/L, respectively.
At the completion of each study, the animal was sacrificed and a necropsy was performed to confirm ECMO cannulae position, and to look for the presence of any cardiac anomalies such as a patent ductus arteriousus (PDA).
All experiments were reviewed and approved by the Animal Care Committee of our institution and all animals received humane care in compliance with the Principle 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 no. 86-23, revised 1985).
Data analysis
Data analysis was performed on a DEC PDP 11/23-L computer (Maynard, MA). Each cardiac cycle was defined using dP/dt. Diastole was considered to begin with the first zero crossing of dP/dt after maximum negative dP/dt and end 40 milliseconds (msec) before maximum positive dP/dt. Beginning ejection was placed at 20 msec after maximum positive dP/dt and end ejection was defined as 40 msec before maximum negative dP/dt. Beat point definitions on all cardiac cycles were inspected and any premature ventricular beats, the subsequent post-extrasystolic beat, and any data collection runs with a greater than 10% change in heart rate were excluded from analysis.
Left ventricular shortening fraction (LVSF) was calculated with the equation:
 |
To define ESPDR, LV end-systolic pressure (Pes) and end-systolic minor axis dimension (Des) were determined for each cardiac cycle that was analyzed during changes in LV preload. Within physiologic ranges of LVP, plotting Pes as a function of Des by linear regression yielded a linear relationship with the equation:
 |
To derive the PRDSW, external LV stroke work (SW) was computed as the integral of LVP and minor axis dimension (D) for each cardiac cycle analyzed during changes in LV preload:
 |
 |
Statistical analysis
All data were expressed as a mean ± 1 standard error of the mean (SEM). Analysis of variance (ANOVA) for repeated measures was used to compare the means of data during and after weaning from ECMO against baseline within each animal. A value of p less than 0.05 was considered to be statistically significant.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
Table 1 shows no significant changes in heart rate and Ded during and after weaning off ECMO when compared to baseline. There is a small but insignificant increase in maximum LVP during ECMO but it decreased to within the baseline range after weaning. Maximum dP/dt showed a gradual but insignificant decrease during ECMO and returned back to baseline after weaning off ECMO. When compared to baseline, both LVSF and VCF, however, showed a significant decrease (p < 0.05) during ECMO but returned to baseline after weaning.
|
|
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
Various investigators [10–12] used the relationship between VCFc and ESS as a load-independent index to assess LV function in infants undergoing ECMO. The discrepancies between the results of these studies may be due to varying clinical conditions. Furthermore, the use of VCFc as a single point measurement of contractility is not valid because the relationship between VCFc and ESS may not always be linear over different contractile states [16].
The goal of this experiment was to measure LV function by load-dependent and load-independent indices in an experimental model of VA ECMO in an effort to clarify the results from previous clinical and experimental studies on this topic. Important features in our protocol included the setting of ECMO flow rate at a fixed percentage of the baseline cardiac output, avoiding the use of inotropic support throughout the study, and maintaining the pO2 in the LA and systemic circulation within physiologic ranges. The animals were not subjected to any ischemic insult prior to the institution of ECMO. In our study, both LVSF and VCF were found to decrease significantly (p < 0.05) during ECMO when compared to baseline. However, the load-insensitive indicators of ventricular contractility, ESPDR and PRDSW, remained unchanged throughout the duration of the study.
End-systolic pressure-volume relationship has been demonstrated to be a load-insensitive and reliable measure of intrinsic cardiac contractility both in isolated hearts [17, 18] and in in vivo models [19]. ESPDR, a modification of the traditional end-systolic pressure-volume relationship in which end-diastolic minor axis dimension instead of volume is used, has been shown to respond in a similar manner to global changes in LV contractility [20]. The relationship between end-systolic minor axis dimension and pressure, ESPDR, is linear and changes in contractility are reflected in the slope (Ees) and x-intercept (Des). The ESPDR in our study had a linear correlation ranging from 0.65 to 0.90. Neither the slope nor the zero-pressure x-intercept showed significant changes during and after ECMO when compared to baseline. Being placed on ECMO alone, therefore, does not have a measurable effect in the intrinsic contractile property of the heart. This conclusion, however, must be made with a certain degree of caution as the linear correlation coefficients for ESPDR were not consistently high due to a wide scattering of data points. Recent experimental studies show that ESPDR can change from a linear to a curvilinear relationship in hearts after being subjected to ischemia or a decrease in LV end-diastolic volume [21, 22].
The relationship between cardiac stroke work and end-diastolic volume, also known as the preload recruitable stroke work, has been demonstrated to be a load-insensitive and highly linear measure of cardiac contractility even after subjecting the LV to a moderate amount of ischemia and reperfusion injury [23, 24]. Furthermore, other experimental studies have shown that LV minor axis dimension is related linearly to ventricular volume [25, 26]. These findings therefore justify our approach of using the PRDSW as a load-independent assessment of cardiac contractility in which end-diastolic minor axis dimension was used to derive PRDSW. The PRDSW in our study had a linear correlation ranging from 0.95 to 0.99. Similar to the ESPDR, no significant differences were found between PRDSW at 4 and 6 hours on ECMO or after weaning off ECMO when compared to baseline. The high degree of linear correlation makes PRDSW a more reliable indicator of cardiac contractility in this model and it appears that being on VA ECMO alone does not have significant deleterious effects on cardiac function. Our conclusions do not contradict the findings of Bavaria and associates [9], where ECMO was shown to cause an increase in the myocardial contractility of post-ischemic hearts. The increase in contractility in their study was thought to be due to augmentation of cardiac output and increased aortic pressure and, hence, an increase in coronary perfusion pressure and blood flow. Our experiment did not show an increase in contractility as our animal’s heart was not subjected to an ischemic insult prior to initiation of ECMO.
Limitations of this study include the brief duration of support during the study. Clinical studies examining the use of echocardiographic indices of function were performed after longer periods of support. The 6-hour endpoint may have been inadequate to allow development of ventricular dysfunction that might become apparent with longer periods of support. Another limitation of study design is the fact that these were normal animals without cardiorespiratory pathology. Children on ECMO often have multifactorial cardiac dysfunction preceding support that might be exacerbated by ECMO. We are currently analyzing a hypoxic animal model of VA ECMO that may detect changes in cardiac function of compromised animals on support.
In spite of these limitations of the study design, the results of our study showed that load-independent indices of ventricular function are unchanged during ECMO support. This contradicts earlier clinical studies that demonstrate deterioration of ventricular function during ECMO. Observed echocardiographic evidence of myocardial dysfunction in neonates on VA ECMO may be a reflection of changes in loading conditions, contractility, ECMO flow rate, pre-ECMO hypoxic and ischemic insults, or other coexisting metabolic conditions. The results of this study imply that children with impaired ventricular function requiring mechanical circulatory support may undergo ECMO without further damage to ventricular contractile function.
|
References |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. M. Stulak, J. A. Dearani, H. M. Burkhart, R. D. Barnes, P. D. Scott, and G. J. Schears ECMO Cannulation Controversies and Complications Seminars in Cardiothoracic and Vascular Anesthesia, September 1, 2009; 13(3): 176 - 182. [Abstract] [PDF]
|
|
|
|
 |
|
 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]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
