Left Ventricular Dynamics and Pulsatile Hemodynamics during Resuscitation of the Fibrillating Heart Using Direct Mechanical Ventricular Actuation
A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy
By
YIRONG ZHOU M.S., Wuhan University, 2011 B. Med., Wuhan University, 2009
______
2018 Wright State University COPYRIGHT BY
YIRONG ZHOU
2018 WRIGHT STATE UNIVERSITY
GRADUATE SCHOOL December 10, 2018 I HEREBY RECOMMEND THAT THE DISSERTATION PREPARED UNDER MY SUPERVISION BY Yirong Zhou ENTITLED Left Ventricular Dynamics and Pulsatile Hemodynamics during Resuscitation of the Fibrillating Heart Using Direct Mechanical Ventricular Actuation BE ACCEPTED IN PARTIAL FULFUILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF Doctor of Philosophy.
______Mark P. Anstadt, M.D. Dissertation Director
______Mill W. Miller, Ph.D. Director, Biomedical Sciences Ph.D. Program
______Barry Milligan, Ph.D. Interim Dean of the Graduate School
Committee on Final Examination
______Mark P. Anstadt, M.D.
______J. Ashot Kozak, Ph.D.
______David Cool, Ph.D.
______Lucile E. Wrenshall, M.D. Ph.D.
______Lynn K. Hartzler, Ph.D. ABSTRACT Zhou, Yirong. Ph.D., Biomedical Sciences Ph.D. Program, Wright State University, 2018. Left Ventricular Dynamics and Pulsatile Hemodynamics during Resuscitation of the Fibrillating Heart Using Direct Mechanical Ventricular Actuation.
The application of mechanical forces during resuscitation of the arrested heart can be used to restore life-sustaining circulation. Open-chest manual massage represents the earliest application of this concept (first described by professor Moritz Schiff in 1874).
Many cardiac compression devices have been developed for cardiac support since that time. Direct mechanical ventricular actuation (DMVA) is a non-blood-contacting device that has demonstrated efficacy of providing both systolic and diastolic support.
The device encompasses the heart and can provide total circulatory support during ventricular fibrillation (VF) or cardiac arrest. DMVA resuscitative support during VF has been shown to be nontraumatic to the myocardium. Notably, resuscitative support using DMVA has the advantage of generating pulsatile flow without blood contact which benefits vital organ perfusion and post-resuscitation neurologic outcome.
However, ventricular and blood flow dynamics during DMVA support have not been well characterized. Specifically, it remains unclear if DMVA support during VF generates ventricular pump function mimicking the native beating heart, or pulsatile flow characteristics similar to the physiological state. The purpose of this dissertation
iv was to better characterize these fundamental aspects of DMVA support during VF.
Experimental data herein demonstrate that DMVA support during VF arrest can result in LV pump function similar to the native beating heart and near-physiological pulsatile flow dynamics. The physiological pulsatile flow generated by DMVA may explain DMVA’s capability for improving resuscitation results. A biventricular mock circulatory system (BMCS) incorporating an anatomical mock ventricle provided supportive in vitro data to further confirm these findings.
v TABLE OF CONTENTS
CHAPTER I: Hypothesis and Specific Aims………………..………………………...1
CHAPTER II: Introduction………………………………………………………...11 Ventricular Fibrillation and Cardiac Arrest………..………………………….11 Resuscitative Mechanical Circulatory Support Devices……………….14 Direct Mechanical Ventricular Actuation………..…………………………..22 Biventricular Mock Circulatory System……..………………….…………….26
CHAPTER III: Materials and Methods...…………………………………………….33
CHAPTER IV: Direct Mechanical Ventricular Actuation during Ventricular Fibrillation Results in Near-Physiological Left Ventricular Myocardial Mechanics……………………………………………………………...... …..55
CHAPTER V: Left Ventricular Diastolic Function Is Returned during Direct Mechanical Ventricular Actuation of the Arrested Heart…………….……....84
CHAPTER VI: Direct Mechanical Ventricular Actuation during Cardiac Arrest Generates Pulsatile Hemodynamics Similar to the Native Beating Heart…...111
CHAPTER VII: Echocardiographic Characteristics of a Mock Ventricle are Similar to the Fibrillating in vivo Ventricle during Direct Mechanical Ventricular Actuation...….……………………………………………….…..138
CHAPTER VIII: Conclusions and Future Directions…………………...... 171
CHAPTER IX: References……….…………………………………………………174
APPENDIX: Commonly Used Abbreviations ...... 197
vi
LIST OF FIGURES
Figure 1. Contemporary resuscitative ventricular mechanical support devices……15
Figure 2. Illustrations of various VA ECMO cannulations……………….……...….. 18
Figure 3. Classification of mechanical circulatory support devices…….….………...20
Figure 4. Schematic of DMVA depicting diastolic expansion (left)
and systolic compression (right)…………………………………………….. 23
Figure 5. Decision tree for DMVA support post SCA…………………..……..……27
Figure 6. The biventricular mock circulatory system (BMCS)…..……………….….29
Figure 7. Sample mock circulation waveforms…………………….……..………….32
Figure 8. Schematic of experimental instrumentation……………….…………..…..35
Figure 9. Four chamber echocardiogram views of three experimental stats...... …..36
Figure 10. Experimental timeline designed to produce VF arrest..……………...…..38
Figure 11. Mathematical relationship among different deformation parameters.....…40
Figure 12. Schematic of circumferential, radial, and longitudinal strain…..……....42
Figure 13. Example volume output from VVI software……………………….….....45
Figure 14. Example strain rate output from VVI software…………………………..46
Figure 15. Image of 3-D printed mold design and the silicone mock ventricle…...…54
Figure 16. Schematic of the DMVA support system (canine study)…………………72
Figure 17. Experimental design. ………………………………………………….…73
Figure 18. Experimental instrumentation…..………………………………..……….74
vii
Figure 19. Four chamber echocardiogram views of (A) native beating heart,
(B) arrested unsupported fibrillating heart, and (C) DMVA supported
VF arrest heart……………………………………………...…………….. 75
Figure 20. LV geometry profiles of native beating heart and DMVA support
during VF arrest. ……………………………………………………...….. 77
Figure 21. LV strain heat maps: (A) regional end-systolic longitudinal
strain (sRLS, %); (B) regional end-diastolic longitudinal
strain (dRLS, %); (C) color scale. ……………………………………….. 80
Figure 22. LV strain rate heat maps: (A) regional peak longitudinal systolic
strain rate (sRLSR, 1/s); (B) regional peak longitudinal diastolic
strain rate (dRLSR, 1/s); (C) color scale.………………………..81
Figure 23. Schematic of the DMVA support system (swine study)………...…….101
Figure 24. Experimental instrumentation………...……...…………………..……102
Figure 25. Four chamber view Echocardiogram Images of (A) native beating
heart (end-diastole), (B) unsupported VF arrested heart, and
(C) DMVA supported VF arrested heart (end-diastole) in swine…….…103
Figure 26. LV geometry of native beating heart and DMVA
support during VF arrest………………………………...……………….105
Figure 27. LV heat maps: (A) regional end-diastolic longitudinal
strain (dRLS, %); (B) regional peak longitudinal diastolic
strain rate (dRLSR, 1/s)………………………….…………………...….108
Figure 28. Regional diastolic strain intra-group comparisons.……………….…109
viii
Figure 29. Regional peak strain rate intra-group comparisons……………………110
Figure 30. Experimental design (pulsatility study)…………………….…………127
Figure 31. Comparisons of all the normalized hemodynamics
and pulsatility measures included in this study…………………………. 131
Figure 32. Aortic power waveforms for three experimental states
at different flow levels. …………………………………….…………… 133
Figure 33. Animal experimental design (mock study)……………………………156
Figure 34. Schematic of the complete mock circulatory system with
biventricular mock heart attached………………………………………..157
Figure 35. Comparable swine and mock ventricles………….……………………160
Figure 36. Mechanical diastolic actuation on the (A) fibrillating swine heart
and (B) mock bi-ventricle. ………………………………………………156
Figure 37. Summary heat map of peak regional longitudinal strains
(RLS, %) at both (A) end-systole and (B) end-diastole
between swine and mock model….……………………………...……… 162
Figure 38. LV GLS and RLS comparison between animal and mock model……164
Figure 39. LV mechanical inotropy and lusitropy reflected by peak strain rate….166
ix
APPENDIX
Figure S1. Regional wall strain intra-group comparisons (canine study)……….....82
Figure S2. Regional strain rate intra-group comparisons……..……...…………….83
Figure S3. Aortic power waveforms of canine and swine………………..……….136
Figure S4. Regional wall strain intra-group comparisons (mock study)…..…..…169
Figure S5. Regional strain rate intra-group comparisons…………………………170
x
LIST OF TABLES
Table 1. Characteristics of resuscitative MCS devices…………………..……….…. 16
Table 2. Summary of parameters to quantify vascular pulsatility…………...………49
Table 3. Basic pressure characteristics of native beating heart
and DMVA support during VF…………………………………………..… 76
Table 4. Left ventricular pump function and myocardial mechanics………….…..…78
Table 5. Hemodynamic characteristics of native beating and
DMVA supported arrested hearts……………………………………….… 104
Table 6. LV diastolic function of native beating and DMVA supported
arrested hearts…………………………………………….…..……………106
Table 7. Estimates of pulsatility used in this study……………………..…………..128
Table 8. Hemodynamic and pulsatility measurements during
baseline, DMVA during VF and ROSC post support…………....……….. 129
Table 9. Echocardiographic and hemodynamic comparison
between swine and mock model………………………………………...…161
xi
APPENDIX
Table S1. Pulsatile flow characteristics of canine and swine at
normal flow level (75%+ baseline flow)………….……..……………...…135
Table S2. Regional longitudinal strain (RLS) comparison
between animal and mock model…………………………………….…..167
Table S3. Regional longitudinal strain rate (RLSR) summary
of animal and mock model…………………………………………….....168
xii
ACKNOWLEDGMENT
I would like to thank the Wright State University Boonshoft School of Medicine and
Graduate School, specifically the Biomedical Sciences Ph.D. program, for giving me this opportunity.
With the sincerest of thanks, I would like to recognize my mentor Dr. Mark Anstadt for providing help, guidance and support, which have been invaluable to my educational, professional, and personal development. It has been a great privilege and honor to work with and learn from you. I would also like to thank the members of my dissertation committee, Dr. Ashot Kozak, Dr. David Cool, Dr. Lucile Wrenshall, Dr.
Lynn Hartzler and Dr. Arthur Pickoff, for their invaluable comments and suggestions on the work contained within this dissertation, and for their willingness to support me and ensure my success. I would like to thank all the past and present members of the
Anstadt Laboratory, especially Benjamin Schmitt, for their intellectual conversations, scientific stimulation, and enjoyable friendship.
Last, but certainly not least a special thank you to my family and friends. I would not be where I am today without any of you. From the bottom of my heart: Thank you!
xiii
This work is dedicated to my parents and my wife. Thanks for everything.
xiv
CHAPTER I HYPOTHESIS & SPECIFIC AIMS
STATEMENT OF THE PROBLEM
Most mechanical circulatory support (MCS) devices currently employed in resuscitation are blood contacting and require cannulation and anticoagulation. They act as direct blood pumps, that typically provide non-pulsatile flow, which can lead to non-ideal tissue perfusion and hematological complications. Alternative direct ventricular compression devices merely augment systolic function, which can further impair diastolic dysfunction that limits their efficacy for resuscitation.
Direct mechanical ventricular actuation (DMVA) under investigation in our laboratory has been demonstrated efficacious for providing resuscitative circulatory support in a fibrillating or non-beating heart (in laboratory and human studies) via directly delivering mechanical forces to the heart surface. DMVA has resulted in favorable cardiovascular and neurologic outcomes following return of spontaneous circulation (ROSC). However, the LV dynamics and hemodynamics during DMVA support of the fibrillating heart have not been well characterized.
The aim of this dissertation was to characterize the effects of optimal DMVA support on ventricular and systemic vascular dynamics during resuscitation of the fibrillating
1
heart. Myocardial mechanics and hemodynamics during DMVA support of the large animal model were collected and comprehensively analyzed. Speckle tracking echocardiography (STE) was utilized as an objective imaging tool to assess and compare the native beating heart and DMVA support in animals. STE provides interrogation of LV wall deformation and synchronous regional dynamics (both mechanical inotropy and lusitropy). Pulsatile flow characteristics were investigated in terms of common pulsatility parameters, harmonic frequency domains, and aortic power waveforms. In addition, a mock circulatory system using silicone mock ventricles was tested as a surrogate to the fibrillating heart during DMVA support.
This was utilized to define potential differences between the fibrillating ventricle and the akinetic silicone material for evaluating ventricular dynamics and pump function during DMVA support.
HYPOTHESIS
Delivering mechanical forces during resuscitation of the fibrillating heart can generate left ventricular dynamics and pulsatile hemodynamics similar to the physiological state.
2
Specific Aim 1: Determine if DMVA resuscitative support during Ventricular
Fibrillation (VF) results in left ventricular pump function similar to that of the
native beating heart (NBH).
Aim1a – Determine if DMVA support of the fibrillating heart results in left ventricular diastolic filling similar to that of the NBH.
Methods
To investigate DMVA’s effect on diastolic filling during ventricular fibrillation, a
VF animal model was utilized in both canine and swine species. Animals were instrumented to collect cardiac output, LV pressures, arterial pressures, and ECHO imaging. LV volumes (EDV, EDVI) and pressures (LVEDP, -dP/dtmax) were compared between baseline and DMVA support, which provide insights of dynamic
LV diastolic pump function.
Aim1b – Determine if DMVA support during VF results in left ventricular systolic pump function similar to that of the NBH.
Methods
The same animal model and experimental timeline were used to assess LV systolic pump function. Parameters were either directly collected or derived from ECHO images and pressure tracings. LV volumes (ESV, ESVI, SV, EF), pressures (LVESP,
+dP/dtmax), and eccentricity (ESDlong/ESDshort) were compared between baseline and
DMVA support.
3
Aim1c – Determine if DMVA support of the fibrillating heart results in global and regional LV wall mechanics similar to that of the NBH.
Methods
The same animal model and experimental timeline were used. 2-D STE were utilized to analyze LV wall dynamics or myocardial tissue deformation. Echo captures consisted of two second B-mode images that include at least 2 cardiac cycles (frame rate is set at 50-60Hz). Global and regional longitudinal strain (GLS and RLS) and strain rate (GLSR and RLSR) was compared between baseline and
DMVA support. Inotropic properties were represented by global and regional systolic strain and strain rate (sGLS, sRLS, sGLSR, and sRLSR, respectively), while lusitropic properties by global and regional diastolic strain and strain rate
(dGLS, dRLS, dGLSR, and dRLSR, respectively).
Aim1d – Determine if DMVA support of the fibrillating heart results in LV inter-regional wall motion relationships similar to those of the NBH.
Methods
Six segments of the LV wall were analyzed using VVI software, which includes basal lateral, mid lateral, apical lateral, basal septal, mid septal, and apical septal wall. The mechanical dyssynchrony index (DI) represents the sum of the wasted energy due to asynchronous contraction and was calculated as the standard deviation of delay in time between peak systolic strain of control region (basal septal wall) and all other regions.
4
Summarized results of all 4 sub-aims (1a-1d)
DMVA support during VF arrest resulted in ejection fraction, cardiac output, global and regional LV myocardial longitudinal strains, and global peak strain rates
(contractility) similar to the NBH (P>0.05) (Table 4). DMVA supported heart also shows equivalent opposing wall delay versus the NBH (P>0.05) (Table 4). DMVA support during VF arrest returned LV end-diastolic pressure, end-diastolic volume, and end-diastolic longitudinal strains to normal values (P>0.05) (Table 6). Global and regional peak strain rates (lusitropy being entirely impacted by direct mechanical forces) were significantly higher during DMVA support (P<0.01) without increase of
LV diastolic dyssynchrony index (P>0.05) (Table 6).
Concluding Statement
DMVA support during VF arrest resulted in near normal systolic and diastolic LV pump function compared to the native beating heart. Furthermore, myocardial mechanical synchrony during DMVA support of the fibrillating heart were similar to the native beating heart. These findings indicate that DMVA support of the fibrillating heart results in myocardial mechanics similar to that occurring during myocardial contraction of the native beating heart. Thus, the mechanically induced inotropic and lusitropic effects of DMVA on the fibrillating LV were similar to that of the native beating heart. This is particularly noteworthy with respect to DMVA’s diastolic actuation as a distinguishing factor when discussing other direct cardiac compression
(DCC) devices that merely provide systolic support. These findings could have
5
important implications on post-resuscitation recovery following return of spontaneous circulation (ROSC). In particular, these findings may have important implications in attenuating diastolic dysfunction following ROSC.
Specific Aim 2: Determine if DMVA resuscitative support of the fibrillating
heart results in pulsatile hemodynamics similar to the physiological state.
Aim 2a – Determine if DMVA support of the fibrillating heart generates near-physiological pulsatile hemodynamics.
Methods
Both canine and swine species were utilized in separate resuscitation protocols.
Animals were anesthetized, underwent sternotomy, and instrumented for hemodynamic monitoring. VF was induced and maintained for 5 mins of circulatory arrest. DMVA was subsequently applied for a 15 min period. Hearts were then defibrillated and allowed to recover. Aortic pressures and flows, time-domain pulsatility measures (PP, EEP, SHE), and frequency-domain flow characteristics
(PI, PPI) were compared between DMVA and baseline. Aortic power waveforms were used as a novel means to more objectively compare pulsatile flow characteristics.
6
Aim 2b – Determine if DMVA support of the fibrillating heart results in pulsatile hemodynamics similar to return of spontaneous circulation (ROSC) at various blood flow dynamic states.
Methods
Canine and swine subjects were anesthetized, underwent sternotomy, and instrumented for hemodynamic monitoring. VF was induced and maintained for 5 mins of circulatory arrest. DMVA was subsequently applied for a 15 min period.
Hearts were then defibrillated and if ROSC achieved allowed to recover. In animals where ROSC was not achieved, DMVA was re-applied. Repeated periods of arrest were used to increase cardiac dysfunction during ROSC. Data was grouped based on percent mean baseline aortic flow: 75%+, 50-75%, and 25-50%. Baseline, DMVA support, and ROSC pulsatile hemodynamics within the three flow groups were compared using previously described pulsatility measures (PP, EEP, SHE, PI, PPI).
Aortic power waveforms were used as a novel means to more objectively compare pulsatile flow characteristics.
Summarized results of both 2 sub-aims (2a-2b)
DMVA produced pulse pressures, surplus hemodynamic energy, and pulsatility index greater than ROSC in all flow groups (Table 8). Furthermore, all pulsatility measures were similar between DMVA in >75% normal flow group versus baseline
(Table 8). DMVA significantly improves peak aortic power relative to
7
post-resuscitation and its waveform inclines in parallel to the native beating heart during early mechanical systole (Figure 32).
Concluding Statement
The results support the hypothesis that DMVA support of the fibrillating heart can result in pulsatile flow similar to the normal physiological state. Notably, results were in large canine and swine species with clinically relevant heart sizes. Data also suggest that the pulsatile flow characteristics resulting from DMVA support of the fibrillating heart were augmented above that of the native beating heart following
ROSC when stratified into hemodynamic sub-groups. The ability to generate physiological pulsatile flow during resuscitative circulatory support has been demonstrated important to cerebral resuscitation and neurological recovery in animal models.
Specific Aim 3: Characterize mock LV mechanics during DMVA support in
contrast to in vivo LV mechanics of the supported fibrillating heart.
Methods
A biventricular mock circulatory system (BMCS) was designed in our laboratory using silicone mock ventricles as surrogates for the akinetic, fibrillating heart. The complete mock circulatory loop was comprised of silicone mock ventricles, atrial reservoirs, and related vascular components for adjusting systemic and pulmonary resistance and flow. Compliance was adjusted by altering the fluid levels in the
8
pertinent reservoirs. Resistance was regulated with adjustable gate valves. The inertance element was accounted for using these adjustable components and taking in to account for fluid viscosity. Instantaneous aortic flow and pressure was recorded. An intracardiac ultrasound probe was integrated in the mock ventricle to collect echocardiographic data. Velocity Vector Imaging (VVI) analysis was used to analyze the mock LV mechanics and wall motion relationships.
Summarized Results
Mock LV echocardiographic images demonstrated LV morphology dynamics similar to the anatomic silhouette and dimensional changes of the fibrillating swine
LV during DMVA support (Figure 36). Ejection fraction, cardiac output, as well as global and regional LV strain were similar in the DMVA supported mock versus swine LV (P > 0.05) (Table 9, Figure 38). Global LV peak strain rate was significantly lower in the BMCS compared to swine ventricle (P < 0.05) (Table 9).
LV regional peak strain rates were significantly reduced in the lateral walls of mock ventricles versus the fibrillating swine ventricle (P < 0.05), while septal regional peak strain rates were similar (P > 0.05) (Figure 39).
Concluding Statement
The results demonstrate that DMVA support of a mock LV can simulate that of the
DMVA supported fibrillating heart. These data help to validate DMVA’s effect on the akinetic, fibrillating heart during resuscitative support. Additionally, the results
9
provide feasibility evidence for using mock silicone ventricles as a reasonable surrogate for studying DMVA’s effects on comparably sized fibrillating hearts.
Therefore, in vitro results can provide indirect supporting evidence for the study hypothesis.
10
CHAPTER II INTRODUCTION
“The heart of creatures is the foundation of life, the Prince of all, the sun of their microcosm, from where all vigor and strength does flow.”
---William Harvey, De Motu Cordis, 1628
Ventricular Fibrillation and Cardiac Arrest
Ventricular Fibrillation (VF) is a severe cardiac rhythm disturbance and the most common immediate cause of sudden cardiac arrest (SCA), which is a major health burden in society and clinical practice (Jalife 2000; Pandit et al., 2013). The incidence of out-of-hospital cardiac arrest (OHCA) reported by the US Resuscitation
Outcomes Consortium (ROC) sites indicates that 140.7 individuals per 100 000
population or 347 322 adults experience cardiac arrest (Mozaffarian D et al., 2016).
Approximately 23% patients have an initial rhythm of VF or ventricular tachycardia
(VT), and the survival to discharge was 32.5%. However, the total survival with treatment post cardiac arrest was merely 12.1% among all adult OHCA patients
(Nichol et al., 2008). The Get With The Guidelines (GWTG)-Resuscitation Registry
reported 20 873 adult patients of in-hospital cardiac arrest (IHCA), with a more optimistic yet low survival to discharge of 24.8% (Merchant et al., 2011).
Undoubtedly, cardiac arrest remains significant complication of many cardiovascular
11
disease states and accounts for more deaths (absolute number) than from breast cancer, lung cancer, or acquired immunodeficiency syndrome (AIDS).
The Heart is the energy origin of the human body. When the heart is diseased, its work is imperfectly performed: the vessels proceeding from the heart become inactive, so that you cannot feel them . . . If the heart trembles, has little power and sinks, the disease is advancing, and death is near (Ebers Papyrus, ~1500 BC). VF is the scenario in which the heart trembles and the vasculature loses flow and pulsatility.
When VF occurs, the heart undergoes turbulent electrical activity, leading to uncoordinated ventricular contraction and no effective pump function. Consequently, hemodynamics suddenly deteriorate, and the loss of vital organ perfusion ensues.
Patients die within minutes due to lack of oxygen delivery to vital organs (especially the brain). The recommended treatment for VF is outlined in Advanced Cardiac Life
Support (ACLS) algorithms (Link et al., 2015). External electrical defibrillation has been proved the most successful treatment for VF and use of early defibrillation via automatic external defibrillators (AEDs) is encouraged (including lay rescuers)
(Hallstrom et al., 2004). However, the success rate can be predicted by the timeliness of intervention based on the Weisfeldt 3-phase time-dependent model (Weisfeldt et al.,
2002). Generally, SCA from VF and pulseless VT undergoes the electric (0-5 minutes), circulatory (5-10 minutes), and metabolic (>10 minutes) phases. Survival to discharge for OHCA can reach 60% for patients treated with rapid defibrillation within the first phase (Patil et al., 2015). Notably, when defibrillation fails to achieve
12
return of spontaneous circulation (ROSC), restoration of blood flow via effective cardiopulmonary resuscitation (CPR) is critical to achieve ROSC (Pantridge et al.,
1966; Halperin et al., 1986). During the third metabolic phase, salvage therapies
(cardiopulmonary bypass (CPB) and extracorporeal membrane oxygenation (ECMO)) plus metabolic drug combinations are required to prevent and correct reperfusion injury and systemic inflammatory disorder (Stub et al., 2015). Notably, implantable cardioverter defibrillators (ICDs) have been used in patients at risk of VF and those who have been resuscitate previously from SCA to prevent recurrence (Yousuf et al.,
2015).
Both AEDs and ICDs are effective for treating arrhythmias, but do not address the underlying causes of SCA, such as ischemic diseases. Mechanical support devices have emerged to improve outcomes of patients suffering SCA refractory to medical management. These devices include modified CPB circuits such as miniaturized
ECMO devices. Other devices including ventricular assist devices (VADs) can be used to bridge patients following resuscitation to recovery from sudden cardiac death.
It is noteworthy that combination therapies are commonly used to treat complex scenarios such as ECMO+VAD, and device plus antiarrhythmics.
According to the 3-phase VF model, rapid restoration of effective organ perfusion predicts ROSC, survival rate and quality of life (QOL) (Tomaselli 2015). A mechanical circulatory support (MCS) device that allows prompt installation and
13
resumption of circulation/perfusion would be ideal for resuscitation. A general review of other MCS devices, focusing on use in resuscitation, will provide a better understanding of both advantages and disadvantages of the MCS devices, and the unique features of direct mechanical ventricular actuation.
Resuscitative Mechanical Circulatory Support Devices
As acute cardiac intensive care continues to improve, SCA patients who require hemodynamic support beyond pharmacological means now have the alternative option of mechanical circulatory support to provide resuscitative life-extending support. Acute MCS devices have evolved over decades such that we now have access to a more nuanced approach to the treatment of cardiac arrest and cardiogenic shock. A variety of commercially available devices and implantation strategies exist, ranging from ECMO to percutaneous temporary VAD (Figure 1), each with unique advantages and disadvantages. The properties of these temporary MCS devices have been compiled into the following table (Table 1).
The ideal MCS device for resuscitation needs to provide hemodynamic support without significant secondary adverse effects. Associated complications may otherwise outweigh the potential therapeutic effect (Ouweneel et al., 2012). In line with the needs for acute MCS, various technical strategies have been developed over the past decades to improve aortic flow and unload the work of critically impaired left ventricle. Notably, mechanical support may be provided to LV alone, right ventricle
14
(D) (D) Direct cardiac compression
;
(A) (A) and (B) are percutaneous left ventricular assist
pport); pport); (E) DMVA, providing resuscitation by active
(ECMO), (ECMO), developed from cardiopulmonary bypass
ventricular mechanical support devices.
resuscitative resuscitative
extracorporeal extracorporeal membrane oxygenation
Figure Figure 1. Contemporary devices; (C) (DCC) devices (represented by Heart Booster, which only provides systolic su al., heart. 2017; Mandawat et Kung al., et from Cohn, 1997; (Adapted expansion 2012). arrested the to and compression
15
2017)
et al., et
Miller Miller
(Adapted (Adapted from
devices.
support
mechanical circulatory mechanicalcirculatory
resuscitative Table 1. Characteristics of of TableCharacteristics 1.
16
(RV) alone, or to both ventricles. Only biventricular assist devices can effectively restore the circulation in the arrested heart. Biventricular support can be combined with an oxygenator (i.e. ECMO) or be utilized alone when underlying pulmonary function is adequate.
ECMO, or other modified extracorporeal life support systems, have been developed from the original cardiopulmonary bypass circuits, the method which is still used today in modern open chest surgeries (Gaffney et al., 2010). ECMO comprises venous-venous (VV) and venous-arterial (VA) configuration. Unlike VV settings
(providing respiratory support only), VA ECMO drains deoxygenated blood from one or more major veins and pumps back oxygenated blood into a major artery, thus providing perfusion of vital organs in the absence of adequate native heart function
(Gilotra et al., 2014). There are peripheral and central cannula options for VA ECLS
(Figure 2). Peripheral cannulation, especially with the femoral vein (inflow to ECMO) and femoral artery (outflow from ECMO), facilitates emergency rescue at the bedside or even portable use (Chung et al., 2014). Central cannulation, on the other hand, has been widely used in pediatric/adult cardiogenic shock and myocardial infarction induced cardiac arrest.
17
Figure 2. Illustrations of various VA ECMO cannulations. A: a femoral vein-femoral artery peripheral cannulation. B: a cannulated carotid artery, a site often used in infants. C: transthoracic right atrial and aortic cardiopulmonary bypass cannulas. (Adapted from Chung et al., 2014)
18
The major concern with blood pumps used for resuscitation is complications related to blood contact. Bleeding and thrombosis remain leading hematological complications during/post ECLS due to the need for anticoagulation. Currently, ECMO features either centrifugal or roller pump and membrane oxygenator. These components cause hematological damage (erythrocytes, platelets, coagulation factors) and generate non-pulsatile flow (Itoh et al., 2016).
Percutaneous ventricular assist devices (pVADs) including the TandemHeart and
Impella and can be used in mechanical cardiopulmonary resuscitation (Gilotra et al.,
2014). These devices aim to unload the arrested heart, and bridge to transplantation or recovery (Sen et al., 2016) (Figure 3). Nonpulsatile pump is utilized in pVADs and complications of VADs have arisen concerning the effects of non-physiologic support, including but not limited to bleeding, acquired von Willebrand Syndrome, renal dysfunction, thromboembolic events, hemolysis, and infection (Travis et al., 2007;
Baric, 2014; Bhimaraj et al., 2015; Cheng et al., 2014; Demirozu et al., 2011; Suarez et al., 2011).
In summary, most mechanical resuscitative devices that are employed clinically to restore circulation and vital organ perfusion (cerebral perfusion being most critical to survival) require direct blood contact, which can cause device thrombosis, thromboembolic events (e.g., stroke), driveline and intravascular infections, bleeding (which is compounded by the need of anticoagulation) and other
19
Figure 3. Classification of mechanical circulatory support devices. Paracorporeal and percutaneous VADs are currently used in the field of resuscitation and discussed in this dissertation. In contrast, intracorporeal VADs are mainly for long-term use in patients with chronic HF (which will not be discussed in detail as they are not utilized for resuscitative circulatory support). (Adapted from Sen et al., 2016)
20
hematological complications. These complications of MCS can lead to significant morbidity and mortality (Jiang et al., 2011). Although MCS devices have been miniaturized as a means for improving their efficacy, these technological advancements have not been associated with any clear evidence of improved resuscitation outcomes.
Another method of circulatory support is direct cardiac compression (DCC). DCC is not a novel concept (Figure 1, Panel E). As a matter of fact, cardiac manual massage has been used to SCA a century ago (Hake, 1874; Bircher et al., 1984;
Alzaga-Fernandez et al., 2005), decades prior to the application of regular CPR
(closed chest compression). DCC devices, involved in open chest cardiopulmonary resuscitation (OC-CPR), continue to apply the concept of cardiac massage using mechanical apparatus. AbioBooster (Kung et al., 1997), CardioSupport System
(Artrip et al. 2000), and HeartPatch (Huang et al. 2003) are representative DCC devices which provide cardiac compression yet lack diastolic support. In fact, these devices have been shown to impair diastolic function. Not surprisingly, more recent development of DCC devices such as the EpicHeart (Moreno et al., 2011a; Moreno et al., 2011b) and the Harvard soft robotic sleeve for biventricular cardiac support
(Roche et al., 2017; Horvath et al., 2018) have focused on assisting the failing heart.
The Harvard device consisting of pneumatic “artificial muscles” is suggested to be better as it supposedly mimics cardiac compression such that there is a twisting effect during compression and an untwisting effect during relaxation. These effects
21
are yet unproven and yet to be illustrated in the literature. Again, these latter DCC devices are not designed to be used in resuscitative support post cardiac arrest.
Alternatively, DMVA provides biventricular compression and expansion (Anstadt et al., 1965; Anstadt et al., 2009) and can rapidly restore adequate organ perfusion in the setting of resuscitative circulatory support during cardiac arrest (Anstadt et al.,
1991a).
Direct Mechanical Ventricular Actuation
Direct mechanical ventricular actuation (DMVA) is a non-blood-contacting ventricular compression device that is clearly distinguished from other DCC devices.
DMVA augments both systolic and diastolic ventricular function when supporting the fibrillating, arrest, or severely failing heart (Anstadt et al., 1995; Perez-Tamayo et al.,
1995; Anstadt et al., 2009). The DMVA cup consists of a semi-rigid outer housing and a flexible inner diaphragm bonded together in the rim and apex. The housing is contoured to fit over the ventricles of the heart in a fashion such that the cup encompasses the ventricles without impinging the atria. A proprietary drive system
(Lifebridge Technologies LLC, Dayton, OH) provides atraumatic systolic and diastolic pneumatic forces that act to compress and dilate the ventricles by inflating and deflating, respectively, the space between the housing and diaphragm (Figure 4).
A continuous low-pressured vacuum maintains atraumatic attachment between the device and the ventricles.
22
Figure 4. Schematic of DMVA depicting (left) diastolic expansion and (right) systolic compression. The pneumatic drive system including vacuum for attachment and pneumatic generation of alternating systolic and diastolic forces (shown on the left side of the diagram).
23
Notably, DMVA was recognized earlier by the American Heart Association (AHA) as having significant potential for resuscitative mechanical circulatory support (Cummins et al., 2003). Adequate perfusion of vital organs in a timely fashion (particularly the brain) is paramount to survival. DMVA’s critical utility in resuscitative circulatory support is attributable to its rapid installation and immediate and sustained biventricular support. Without the need for cannulation (necessary for ECMO and pVADs), it allows for documented installation times of less than 5 minutes in both animal (Brown et al., 1989; Griffith et al., 1992) and clinical (Lowe et al., 1999;
Anstadt et al., 1991b) studies. Furthermore, DMVA provides pulsatile flow, further facilitating cerebral and renal perfusion post SCA with significant improvements of cerebral oxygen consumption, numbers of neurons, and ischemic damages. Both laboratory and clinical investigations have indicated DMVA’s superiority in cardiac and cerebral resuscitation when compared to closed-chest compression (CC-CPR), open-chest cardiac massage (OC-CPR), and full cardiopulmonary bypass (CPB)
(Anstadt et al.,1991a;1991d;1992;1993). Pulsatile flow during DMVA support was felt to be of critical importance to the superior cerebral resuscitation results when compared to conventional cardiopulmonary bypass. Notably, in the clinical setting,
DMVA could likely to installed more rapidly than any existing form of MCS and/or
CPB. However, these experimental studies biased CPB results by instituting support at similar arrest times. Therefore, the additional benefit DMVA might has on resuscitation results due to rapid installation should be considered.
24
In addition, non-pulsatile flow resuscitative devices may cause secondary complications as has been observed in longer-term use of such devices. For example, gastrointestinal (GI) bleeding may result from inadequent perfusion which is thought to be related to non-physiological, non-pulsatile flow. Furthermore, MCS is associated with impaired clotting and bleeding associated with the need for anticoagulants to prevent thromboembolic events (Wever-Pinzon et al., 2013;
Aggarwal et al., 2012). Renal dysfunction and compromised lymphatic flow have also been observed with non-pulsatile flow. The absence of the pulsatile circulation reduces vascular bed patency and shear stress, thereby leading to endothelial dysfunction (O'Neil et al., 2012). DMVA resuscitative support potentially circumvents these consequences since the device can generate pulsatile reperfusion.
Importantly, DMVA is non-blood-contacting, which circumvents various hematological complications (device thrombosis, internal bleeding, stroke, infection, etc.) without the requirement of anticoagulation. Distinguished from non-blood contacting DCC devices (merely provide systolic support), DMVA also augments diastolic filling. Previous studies have demonstrated significant improvements in the left ventricle (LV) diastolic -dP/dt (negative change in pressure/time) and diastolic myocardial strain rates (change in myocyte length/time) during DVMA support
(Wozniak et al., 2007a;2007b; Anstadt et al., 2008;2009). Efficacious diastolic support may relate to restoration of normal coronary circulation, which potentially increases tolerance to ischemia post cardiac arrest (Anstadt et al., 1995).
25
There is compelling evidence that DMVA does not traumatize the heart in use for resuscitative circulatory support for hours to days (Anstadt et al., 1987; 1990; 1992).
In addition, anecdotal clinical evidence revealed that the heart can be supported for days to months without traumatizing the myocardium (Lowe et al., 1991;1999). The proprietary drive system has been developed to ensure atraumatic and efficacious pneumatic forces applied to the ventricles. Surgical expertise and appropriate monitoring are required for optimal DMVA support. Cup size selection and avoiding complications associated with technical mishaps during thoracotomy must also be considered (Anstadt et al., 1991a). However, use of DMVA has unique advantages for resuscitation compared to other MCS devices. Once applied, DMVA has the further capability of acting as bridge to decision following successful resuscitative circulatory support (Figure 5).
Biventricular Mock Circulatory System
Assessment of mechanical circulatory support devices in vitro using a mock circulation loop (MCL) is frequently utilized before extensive in vivo testing.
Successful development of an MCL that reasonably simulates the physiology of the cardiovascular system requires a pertinent knowledge of cardiovascular anatomy and hemodynamics. Such systems can promote the evaluation of the device and its capability to treat pathological conditions.
26
Figure 5. Decision tree for DMVA support in the setting of SCA and uncertain neurologic status. DMVA support after VF induced cardiac arrest provides hemodynamic stabilization (near-physiologically), which allows time for assessment of organ perfusion and neurologic status of the patient. If the SCA patient achieves both neurologic and cardiac recovery, DMVA support can serve as a bridge to recovery and subsequently be weaned off. If the patient does not recover neurologic function, withdrawal of DMVA support is required. If the patient awakens but does not have adequate myocardial function, DMVA support allows time to assess indications for heart transplant (bridge to transplantation, BTT), VADs (bridge to
VADs) or destination therapy (DT). The neurologically-recovered patient, yet disqualified from a heart transplant or VAD, presents a dilemma that might be described as a bridge to nowhere. (Adapted from Chung et al., 2014)
27
Various MCLs have been developed to provide test beds for different types of support devices. The complete mock circulation system should be a circuit consisting of cardiac and vascular components for both the systemic and pulmonary circulations
(Timms et al., 2005). These components are then assigned mechanical properties obtained from the natural cardiovascular system to reproduce desired hemodynamic characteristics.
Our laboratory has been developing a biventricular mock circulatory system (BMCS) to evaluate DCC devices in vitro. As a complete MCL, the BMCS platform consists of a heart (mock bi-ventricles and atrial reservoirs connected by atrioventricular (AV) platform) and vasculature (systemic/pulmonary resistance and compliance elements)
(Figure 6). Using anatomical measurements from extracted swine hearts, a
SOLIDWORKS® computational model is developed to 3D print a ventricular mold.
Silicone rubber mock ventricles (including intact left and right ventricles and septal wall) are fabricated from this 3D-printed mold (Gregory et al., 2009). This custom-designed biventricular mock model efficiently simulates the fibrillating heart when supported by DMVA. The mock ventricles are integrated with the AV platform which is attached to two atrial reservoirs. The pulmonary valve and tricuspid valve are seated in the base of the AV platform providing minimal dead space between the valves and the right ventricle (RV). The aortic and mitral valves are positioned above the AV platform. Silicone mock ventricles are firmly clamped over the inner mounting flange of the platform.
28
m is shown on the left. The The is m left. the on shown
C512syste ECHO
.
(BMCS)
ystem
s
irculatory irculatory
c
ock ock
biventricular m biventricular
ventricle is attached to the atrioventricular (AV) platform and displayed for details displayed and on for right. the the to (AV) attached is platform atrioventricular ventricle
-
The
. .
6
Figure Figure bi silicone
29
Vasculatures of systemic and pulmonary circulation are based on the four-element
Windkessel model, consisting of aortic valve resistance (characteristic impedance), peripheral resistance, total arterial compliance, and aortic inertance (Timms et al.,
2011; Segers et al., 2008). Compliance was adjusted by altering the fluid levels in the pertinent reservoirs. Resistance was regulated with adjustable gate valves. The inertance element was accounted for integrating these adjustable components and taking in to account for fluid viscosity. Instantaneous aortic flow and pressure was recorded (Figure 7). An intracardiac ultrasound probe was integrated in the mock ventricle to collect echocardiographic data. Velocity vector imaging analysis was utilized to analyze the mock LV mechanics and wall motion relationships.
In summary, the BMCS mock system can be utilized to test the effects of DMVA on the non-beating, asystolic mock ventricle. In this manner, the effective DMVA can further be tested and compare to that found in the native, fibrillating heart. Animal in vivo models were utilized in specific aim 1 and specific aim 2. In specific aim 3, mock in vitro model was used to indirectly test the hypothesis since fibrillating ventricle and mock ventricle share the same akinetic feature without DMVA support.
Silicone, given its compliance and stretch properties similar to myocardial tissues, could be used to fabricate an eligible surrogate of the fibrillating heart (Gregory et al., 2009). The study could shed light on data collected from the animal model since
DMVA is actuating an akinetic mass without any of the compensation found in heart tissue.
30
Mock ventricles were designed to display similar dynamic properties compared to the fibrillating heart during DMVA support, including echocardiographic
(dimensions, deformation), and hemodynamic comparisons (aortic pressure and flow). Major parameters should be consistent with animal data and thus prove the efficacy of this mock ventricle. Discrepancies could be observed in certain parameters due to imperfect simulation. Strain data might be confounding since silicone cannot entirely replicate myocardial behavioral properties. However, the parallel of systolic and diastolic displacement should be observed. Data could be used as feedbacks to optimize the design of mock ventricles (hardness, volume, wall thickness). One major advantage of our BMCS is that it allows for prompt modulations in repetitive tests.
31
Figure 7. Sample mock circulation waveforms showing near-physiological patterns of arterial flow and pressure. Red curves show systemic hemodynamics and blue curves display pulmonary hemodynamics.
32
CHAPTER III MATERIALS AND METHODS
Surgical protocol
All the animals received humane care in compliance with the Guide for the Care and
Use of Laboratory Animals, published by the National Institutes of Health. All surgeries were under the regulation of the Laboratory Animal Resources (Animal
Use Protocol #885) at Wright State University.
The large animal models were comprised of mixed breed canine (n=20, 37±5.9kg) and Yorkshire Swine (n=10, 85±5.1kg). These two models were used to provide a wider variety of heart sizes comparable to adult human. Subjects were pre-anesthetized with Telazol 6 mg/kg, Xylazine 3 mg/kg, and Atropine 0.02 mg/kg, followed by tracheal intubation and mechanical ventilation. Anesthesia was maintained with continuous 1-2% isoflurane throughout the experiment. Bilateral femoral cutdowns were used as vascular access for drug infusion. Phenylephrine, a selective α1-adrenergic receptor agonist, was used to maintain arterial pressure for adequate perfusion among the arrest/recovery cycles. Surface electrocardiogram
(ECG), hemodynamics, pulse oximetry, end-tidal carbon dioxide, and rectal temperature were monitored using a computer-based acquisition system.
33
Subjects were instrumented with pressure transducers (Spectramed P23XL) in the
LV, ascending aorta, and superior vena cava (Figure 8). An ultrasound flow probe
(Transonic COfidence) attached to a perivascular module (TS420) was placed around the ascending aorta to capture instantaneous flow data, which is computed using the difference in upstream and downstream transit time. A 10-MHz Acuson
AcuNav10F ultrasound catheter was positioned through the right jugular vein for intracardiac imaging. Four-chamber views were obtained over a couple heart cycles on the Acuson Sequoia C512 ECHO system (Figure 9). A median sternotomy and pericardiotomy was performed to expose the heart and install both the flow probe and DMVA cup. The cup housing encompassed the greatest surface area of the ventricles without impinging the atria.
Considerations of the Resuscitation Model
The large animal model utilized in this project is a standard cardiac arrest-resuscitation model (Vognsen M et al., 2017; Cherry et al., 2015). VF Cardiac arrest was electrically-induced in a health animal subject without any pre-existing cardiovascular or other systemic diseases. Notably, experimental models typically utilize normal animal subjects and may have other significant differences from clinically relevant circumstances surrounding cardiac arrest and resuscitation.
Clinically, the underlying cause of out-of-hospital cardiac arrest varies by age group.
Approximately 5% to 10% of cardiac arrest cases occur in the absence of structural heart diseases (Chelly et al., 2012). However, a large proportion of patients with
34
and
, ,
pressure
instrumented for aortic and for aortic instrumented flow
placed down the left jugular vein to acquire
ere
was
instrumented for pulsatile and apical vacuum apical pressures. for pulsatileinstrumented and
Animals w Animals
catheter
also also
was
ultrasound
mental instrumentation. mental
n n intravascular
pressure. A
Schematic of Schematic experi
Figure 8. Figure left ventricular DMVA echocardiography. cup The intracardiac
35
LV: LV:
. .
s
model
and and swine
ventricular ventricular fibrillating
), ),
canine canine
and D
both
acquired from acquired
) )
native native beating heart (A
and F
the the
), ), and supportedDMVA VF heart (C
Four Four chamber echocardiogram views of
and E
Figure Figure 9. heart (B fibrillation. VF: RV: ventricular ventricle; right ventricle; left
36
cardiac arrest is elderly and has chronic systemic disorders such as coronary atherosclerosis, hypertension, congestive heart failure, diabetes mellitus, emphysema or renal disease. Furthermore, patients with cardiac arrest are not typically under anesthesia when they unanticipatedly experience sudden cardiac arrest. These discrepancies between the clinical setting and the experimental models used in this dissertation should be understood. However, these discrepancies should not have altered the experimental results pertaining to the hypothesis being tested. Notably, other important clinically relevant variables related to cardiovascular disease and cardiac remodeling would deserve further study if these results can be translated to all such circumstances.
Experimental Timeline
See detailed flowchart in Figure 10. Post-sternotomy baseline was collected as physiological level in this study. Surgical skills were trained in preliminary experiments. Any case with major bleeding or pericardial laceration was be excluded. After baseline data collection, the heart was induced VF via a 9-volt battery applied to the epicardial surface. VF/SCA data was collected for a period of
5 minutes. The DMVA cup was installed on the fibrillating heart (Immediately below atrioventricular groove) to assess dynamics during minimal intrinsic cardiac function. After a 15-minute support period, attempts were made to defibrillate the heart. If defibrillation was unsuccessful, DMVA was re-applied on the heart for another round of support. If the sinus rhythm was restored (there is also the
37
Figure 10. Experimental timeline designed to produce ventricular fibrillation
(VF). This design enables repeated resuscitative cycles to test DMVA during VF for up to 4 hours. In specific aim 1, baseline native beating heart (blue) and DMVA support during VF (green) were analyzed. In specific aim 2, baseline native beating heart (blue), DMVA support during VF (green), and ROSC (red) were analyzed.
The design allows for evaluation of DMVA support and ROSC at different flow levels. In specific aim 3, DMVA support of the fibrillating heart (red) and DMVA support of the mock ventricles were analyzed.
38
possibility of spontaneous mechanical defibrillation), cardiac output as a percentage of baseline was used to define the extent of recovery. This design ensures at least
4-5 arrest/recovery cycles to test DMVA during VF, allowing for evaluation of
DMVA support and ROSC at different flow levels addressed in specific aim 2.
2-D Speckle Tracking Echocardiography
Speckle tracking echocardiography (STE) is a valuable, objective imaging modality utilized in our laboratory as well as by many other research facilities for characterizing ventricular morphology, cardiac function, and complex myocardial strain dynamics. Superior to Tissue Doppler Imaging (TDI), the other commonly used imaging approach, STE not only provides less time-intensive offline calculation of myocardial velocities and deformation indices as strain and strain rate, but also has favorable signal-to-noise ratio and frame rate, and most importantly, angle independence (Blessberger et al., 2010). Both TDI and STE could analyze deformation parameters such as velocity, strain, and strain rate, and these three can be mathematically converted (Figure 11).
Echo captures consist of two second B-mode images that include at least 2 cardiac cycles (frame rate was set at 50Hz). ECG signals were collected as part of the STE system but are considered unreliable for synchrony study due to mechanical noise during DMVA support. Offline analyses were performed using Siemens workplace software SYNGO (Siemens Medical Solutions, Mountain View, CA) and the
39
Figure 11. Mathematical relationship among different deformation parameters and mode of calculation for speckle tracking echocardiography (STE) and tissue Doppler imaging (TDI). STE primarily assesses myocardial displacement
(depicted in blue squares), whereas TDI assesses tissue velocity (depicted in orange squares). (Adapted from Blessberger et al., 2010)
40
included module --- Velocity Vector Imaging (VVI) to track cardiac wall motion.
The endocardial border of LV was traced on the end-diastolic frame (up to the level of the valve annulus). Experiences with tracings indicated that cardiac cycles should be individually analyzed to generate high-fidelity strain curves. Details regarding
ECHO settings and operation are included in Chapter IV.
Speckle tracking technology defines a cluster of speckles (acoustic markers/reflections from micro-boundaries in the tissue) and follows this cluster frame to frame (Figure 12, right). Speckle tracking can be utilized to assess and interrogate complex ventricular and/or myocardial dynamics. These types of interrogation can be expressed in terms of strain. Strain is typically expressed in terms of changes from initial/original length (Lagrangian strain). Alternatively, it can be described as change in instantaneous length from point of prior reference
(natural strain). Lagrangian strain is typically utilized in mainstream cardiac ultrasound systems to characterize ventricular and myocardial function and dynamics. Lagrangian strain can be calculated as (Voigt et al.,
2015). These measures can either represent discrete strain endpoints (e.g., end-systole and end-diastole), or express rate of strain change (i.e. strain rate).
Notably, peak myocardial strain rate is a relatively load-independent measure of myocardial function. Strain and strain rate can both be measured along longitudinal, radial and circumferential axes (Figure 12, left). All three planes are interrelated, assuming myocardial tissue volume remains constant during the cardiac cycle. Only
41
Figure 12. (Left) Layer-specific evaluation of circumferential, radial, and longitudinal strain. (Right) Example of longitudinal strain calculation.
Longitudinal strain can be computed by , L is the tissue length.
(Adapted from Gupta et al., 2015; Blessberger et al., 2010)
42
longitudinal parameters were analyzed because the air-filled DMVA cup tended to disrupt the ultrasound beam in the short-axis view. Strain rate (systolic and diastolic) was derived from Lagrangian strain to indicate myocardial compression and recoil properties. Regional strain (six segments on apical view) data were used to evaluate regional LV function and synchronous wall motion relationships. Strain and its temporal derivative strain rate are dynamic measurements and can reflect intrinsic myocardial forces during both systole and diastole (i.e. inotropy and lusitropy)
(Geyer et al., 2010; Voigt et al., 2015).
Speckle tracking allows relatively load-independent measures of global and regional myocardial function and effective method for assessing pertinent changes in LV volumes (Geyer et al., 2010; Leung et al., 2010; Smiseth et al., 2016). Measurement of global longitudinal strain (GLS) by STE has emerged as a sensitive index of cardiac and myocardial performance. Other echocardiographic indices of cardiac function such as ejection fraction (EF) and myocardial function such as regional shortening are less precise and clearly load-dependent. GLS, as commonly used clinically, refers to systolic global longitudinal strain and will be termed sGLS in this dissertation to differentiate with diastolic global longitudinal strain (dGLS).
Regional contractile function was efficiently assessed by regional strain profile.
Admittedly, both strain and strain rate are still partially dependent on loading conditions. Peak systolic strain rate, however, being an early systolic event, correlates more with contractility than EF and GLS. Meanwhile, global and regional
43
peak diastolic strain rate (dRLSR) occurs during early diastole (relaxation) and thus is an important component of diastolic function. Characterizing all these aspects of diastole can be accomplished using speckle tracking. This method can be used to interrogate native beating heart as well as the DMVA supported heart during ventricular fibrillation. Therefore, speckle tracking is chosen as an objective means of defining diastolic function and understanding the unique perspectives of mechanical/myocardial and ventricular lusitropy in this study (Hamlin et al., 2012;
Flachskampf et al., 2015).
Figure 13 displays an example of LV volume analysis using VVI module (Siemens
Medical Solutions, Mountain View, CA). The software allows computation of segmental and global volumes and eccentricity parameters (long/short-axis diameter). Figure 14 shows a representative screenshot of strain rate data derived from traced ECHO images. Synchrony parameters as opposing wall delay and dyssynchrony index were also computed from this data set.
44
Figure 13. Example output from VVI module displaying global and segmental volumes. Left ventricle is divided into six segments in the offline analysis, segments which are basal lateral, mid lateral, apical lateral, basal septal, mid septal, and apical septal wall. Top left is the condensed ECHO image. Bottom left is the segmental ejection fraction. Upper right is the global volume (red curve) and dV/dt (blue curve) overtime. Middle right is the long-axis dimension (red curve) and short-axis dimension (blue curve) overtime. Lower right is the segmental volumes overtime.
Green curve represents the recorded ECG signal.
45
Figure 14. Example output from VVI module displaying global and regional strain rates. Left ventricle is divided into six segments in the offline analysis, segments which are basal lateral, mid lateral, apical lateral, basal septal, mid septal, and apical septal wall. Values of regional longitudinal strain rate (RLSR) are shown on the top table, as well as the time to peak strain rate. Global longitudinal strain
(GLS) is calculated by averaging six regional strains., while global longitudinal strain rate (GLSR) is computed by averaging six regional strain rates. Waveforms of
GLSR and RLSR overtime are displayed in the bottom right panel. On the right are two arches showing regional time to peak and phase.
46
Pulsatility and Pulsatile Flow Assessment
The natural beating heart epitomizes optimal cardiovascular hemodynamics, which consist of cyclic/pulsatile blood pressure and flow patterns. Because pressure and flow are intricately connected in this context, the integrated pulsatile hemodynamics can generally be referred to as pulsatility (Soucy et al., 2013). Instant aortic flows and pressures were simultaneously recorded during experiments, as the basic parameters defining the blood flow dynamics. In-depth analysis of these instantaneous waveforms of the two were performed in this study, and series of parameters (in both time and frequency domains) were calculated to describe vascular pulsatility. All metrics used in this study are listed in Table 2. Arterial pulse pressure and pulsatility index are the first common indices to describe pulsatility. Pulse pressure (PP) is the difference between the maximum and minimum pressure, while pulsatility index (PI) is the difference between peak systolic and diastolic blood flow velocity, divided by the mean velocity over the cardiac cycle. The latter is not used in this study since the flow velocity cannot be directly measured during the experiment.
Importantly, pulsatile flow depends on the energy gradient, rather than the pressure gradient (Undar et al., 2002). By mathematical modeling, it has been demonstrated that at the same mean arterial pressure pulsatile flow provided 2.4 times as much energy as a nonpulsatile flow. Energy equivalent pressure (EEP), first defined by
Shepard et al in 1966 (Shepard et al., 1966), is a flow-weighted pressure metric able
47
to incorporate more hemodynamic information. EEP is calculated as the ratio between the area under the curve of power and the area under the flow curve. Undar et al have reported extensively on the merits of EEP, and proposed surplus hemodynamic energy (SHE), calculated by multiplying the difference between EEP and mean arterial pressure (MAP) by 1332, as a novel method for precise quantification of different levels of pulsatility (Undar, 2005). SHE precisely quantifies the extra energy required for production of pulsatile flow in terms of energy (not pressure) units and is thus a physiologically relevant measure of pulsatility (Souncy et al., 2013).
Flow waveform was analyzed in the frequency domain, using a fast Fourier transform algorithm and m-files programmed in MATLAB. Pulsatility index (PIfreq) and pulse power index (PPI) were measured in the frequency-domain. The PI is the sum of the squares of its harmonic components divided by the square of the DC component (mean flow), while the PPI integrates frequency value (ω) to better quantify the power of a pulsatile waveform versus a nonpulsatile equivalent flow
(Kaebnick et al., 2007). Furthermore, waveform of aortic power (product of simultaneous aortic flow and pressure) was analyzed in the time domain.
Waveforms were broken into individual cycles and standardized to a representative
0.7 second cycle, allowing for comparisons between experimental states. Details are shown in Chapter VI.
48
Table 2. Summary of Parameters to Quantify Vascular Pulsatility. Ai is amplitude of i’th harmonic of flow, A0 is amplitude of mean flow and ω is angular frequency of flow.
Measure Equation
Pulse Pressure PP = Psys - Pdias [mmHg]
Energy Equivalent ∫ 푓푙표푤 ∗ 푝푟푒푠푠푢푟푒 푑푡 퐸퐸푃 = [mmHg] ∫ 푓푙표푤 푑푡 Pressure
Surplus Hemodynamic 푆퐻퐸 = 1332 ∗ (퐸퐸푃 − 푀퐴푃) [erg/cm3] Energy
2 Σ(퐴푖 ) Pulsatility Index 푃퐼푓푟푒푞 = 2 [unitless] 퐴0
2 2 Σ(휔푖 )(퐴푖 ) Pulsatility Power Index 푃푃퐼 = 2 [unitless] 퐴0
49
Synchronous Wall Motion and Function Analysis
Given that ECG recordings are noisy during DMVA support, the analysis used ventricular strain as landmarks for calculating wall delay (Tε) and the dyssynchrony index (standard deviation of Tε) (Yu et al., 2002; Lim et al., 2008; Liang et al., 2013).
The level of inter-regional wall motion relationship in LV imposed by DMVA was evaluated. CRT (cardiac resynchronization therapy) has been a considerable advance in the therapy of Chronic HF, and global dyssynchrony index has been used as a strong predictor of reverse remodeling after CRT (Yu et al., 2004). These same measures can be applied for assessing myocardial passive regional motion during
DMVA support and developing a tool to predict long-term survival in future animal and human trials. Clinically, dyssynchrony in heart diseases includes atrioventricular, interventricular and intraventricular components. In this dissertation, intra-LV synchronous regional function was analyzed between the native beating heart and
DMVA support of the fibrillating heart.
Intraventricular dyssynchronous regional function/wall motion compromises effectiveness of LV contraction and leads to waste of myocardial energy. The dyssynchrony index (DI) represents the sum of the wasted energy and is computed by the equation at the end of next paragraph. However, its accuracy depends on the precision of the method for measuring regional wall motion. It has been shown that regional peak longitudinal strain collected by STE provides an accurate qualification
50
of regional wall motion, with global longitudinal strain (GLS) strongly correlated to
LV ejection fraction.
Six segments of the LV wall were analyzed using VVI software, which includes basal free wall, mid free wall, apical free wall, basal septal wall, mid septal wall, and apical septal wall. One concern would be insufficient resolution necessary to detect differences, since most clinical assessments of synchrony use
16,17,18-segment models by analyzing all three ultrasound planes (Lim et al., 2008).
Alternatively, we characterized deformation of 50 constituent points (50 potential small regions) of the LV wall, and calculate a more comprehensive global DI for the study. The equation is shown as follows.
Mock Ventricle Test
A SOLIDWORKSTM (Dassault Systèmes, France) computational model was developed based on anatomical measurements of an appropriately sized swine heart.
This model consists of an outer shell to define the external ventricular surface and an insert to define the shape of the ventricular cavities. The two pieces of this model were fabricated using a 3-D printer (Makerbot Replicator 2, MakerBot® Industries,
LLC, NY). The mock ventricles were then constructed by pouring silicone
51
(Ecoflex® Supersoft Silicone, Smooth-On, Inc, PA) in between the two molds
(Figure 15). The physiologic ventricles have been estimated to possess a shore-hardness around OO-15 (Gregory et al., 2009). Therefore, a mixture of
OO-10 and OO-20 silicones were used to create intermediate values approximating that observed in the native heart. Silicone has been tested to be the optimal material for constructing a biventricular heart model due to its ability to endure prolonged cyclic mechanical stress from the DMVA cup. Attached to our BMCS to simulate physiologic loads, the mock ventricle can be used to assess DMVA in a more controllable manner than the fibrillating heart.
An atrial reservoir was designed to maintain the constant volume in the BMCS throughout the experiment. The volume of water in this reservoir, as well as the preload, may be varied by changing the level of a runoff port. Gate valve, located between the compliance chamber and the atrial reservoir, was used to adjust the level of arterial resistance (quantified as number of turns). The valve at the top of the compliance chamber was designed to connect the large syringe sucking air out of the chamber. Consequently, the level of the water/air interface represents the desired compliance that is seen in the vasculature (quantified as the scale readings on the side of the chamber). Real-time flow and pressure data were recorded using
LABVIEW programming (at a sampling rate of 200 Hz).
52
Comparable level of flow and pressure were tested in the BMCS during DMVA support. The integrated mock ventricles were utilized to simulate the DMVA supported fibrillating heart and to analyzed LV dynamics and wall motion relationships as well as chamber dimensions.
53
Figure 15. Image of 3-D printed mold design and the silicone mock ventricles.
Ventricle shell (A) and inserts (B) are shown in a 3-D composition. Top (C) and side view (D) of the mock ventricle are depicting the close structure of a native heart.
54
CHAPTER IV Direct Mechanical Ventricular Actuation during Ventricular Fibrillation Results in Near-Physiological Left Ventricular Myocardial Mechanics
This chapter addresses SA 1 of this dissertation.
Specific Aim 1: Determine if DMVA resuscitative support during Ventricular
Fibrillation (VF) results in left ventricular pump function similar to that of the native beating heart (NBH).
Aim1a – Determine if DMVA support of the fibrillating heart results in left ventricular diastolic filling similar to that of the NBH.
Aim1b – Determine if DMVA support during VF results in left ventricular systolic pump function similar to that of the NBH.
Aim1c – Determine if DMVA support of the fibrillating heart results in global and regional LV wall mechanics similar to that of the NBH.
Aim1d – Determine if DMVA support of the fibrillating heart results in LV inter-regional wall motion relationships similar to those of the NBH.
55
RATIONALE
Ventricular Fibrillation (VF) is the most severe cardiac rhythm disturbance and the major immediate cause of cardiac arrest (CA) (Lang et al., 2006). VF induced cardiac arrest is a global health burden in society and clinical practice, from which the total survival rate is merely 12.1% and 24.8% for out-of-hospital and in-hospital
CA patients in the United States (Mozaffarian D et al., 2016). When VF occurs, the heart trembles and blood flow loses its pulsatility. Specifically, the heart undergoes turbulent electrical activity and consequent uncoordinated ventricular contraction, which leads to sudden and extreme drop of the hemodynamics. Without prompt treatment, patients would either die within less than 10 minutes or suffer from long-term hypoxic ischemic brain injury due to lack of adequate oxygen delivery.
Survival to discharge can reach 60% for CA patients treated with rapid defibrillation within the first 5 minutes when VF initiates, which is termed as the electric phase
(Weisfeldt et al., 2002; Patil et al., 2015). If defibrillation fails to achieve return of spontaneous circulation (ROSC), patient with VF enters the second circulatory phase (5-10 minutes). Cardiopulmonary resuscitation (CPR) is necessarily applied to restore blood flow and perfusion of vital organs including myocardial perfusion itself. During the third metabolic phase (VF sustains for over 10 minutes), acute mechanical circulatory support (MCS) devices such as extracorporeal membrane oxygenation (ECMO) plus metabolic drug combinations have emerged in practice to
56
improve neurologic outcomes of patients suffering CA refractory to medical management and defibrillator therapy (Siao et al., 2015).
Direct Mechanical Ventricle Actuation (DMVA) is a non-blood-contacting ventricular compression device that augments both systolic and diastolic function to support the failing or fibrillating heart (Anstadt et al., 1995; Perez-Tamayo et al.,
1995; Anstadt et al., 2008; Zhou et al., 2018b). DMVA can provide mechanical forces to the heart surface and pump the heart in a physiological fashion in terms of ventricular compression and expansion (Figure 16). Previous investigations have evaluated left ventricular (LV) myocardial strain rates and function during DMVA support of the failing heart (Anstadt et al., 2009). However, DMVA’s effects on myocardial mechanics in the fibrillating ventricle have not been well characterized.
The purpose of this study was to determine if the application of DMVA to the fibrillating heart generates LV pump function and myocardial mechanics similar to the physiological state in a mature, clinically-relevant animal model.
RESEARCH DESIGN
The Wright State University Lab Animal Care and Use Committee approved the experimental protocol, and all animals were treated in compliance with the Guide for the Care and Use of Laboratory Animals of the National Academies, published by the
National Institutes of Health. All surgeries were performed by a qualified, trained,
57
and experienced cardiothoracic surgeon team.
Ventricular Fibrillation Model
Mature mixed-breed canine (n=12, 30±2.6kg) were studied using a model of repeated periods of cardiac arrest (Figure 17). Animals were anesthetized (1-2% isoflurane), underwent median sternotomy and pericardiotomy, and instrumented for hemodynamic and intracardiac echocardiography (Figure 18) (Chapter III, surgical protocol). Baseline data was collected of the native beating heart (NBH). VF arrest was then induced using a 9-volt electrical shock for 5 mins of unsupported circulatory arrest. DMVA was then applied during VF for a 15 min followed by defibrillation attempts. Other than the administration of standard doses of lidocaine, no other cardiovascular agents such as vasopressors or inotropes were administered throughout the experiment. The whole resuscitation cycle was repeated for approximately 4 hours. Animals were euthanized by potassium chloride injection
(1-2 mmol/kg) under anesthesia at the end of the experiment (Beaver et al., 2001).
DMVA support rates were adjusted to 120 times per minute, which matches natural rhythms in canine subjects. DMVA cup size was set at 85-90 mm diameter to ensure a good fit at the rim of the cup near the atrioventricular groove).
Hemodynamic Data Collection
Cardiac output (CO) was measured by a TS420 perivascular flowmeter (Transonic
Systems Inc, Ithaca, NY) connected to an ultrasound probe placed on the ascending
58
aorta. Aortic and LV pressure were monitored through two identical P23XL transducers (Spectramed Inc, Mt Vernon, OH) introduced into the aorta and LV via the right carotid respectively. Flow and pressure data were then periodically collected throughout the experiments for 10 sec intervals at 200 Hz utilizing
LabVIEW data acquisition software (National Instruments, Austin, TX).
2-D Speckle Tracking Echocardiography
Intracardiac echocardiography was interrogated on the canine VF model using an
ACUSON Sequoia C512 System (Siemens Medical Solutions, Mountain View,
CA). A 10-MHz ACUSON AcuNav10F intravascular catheter (Siemens Medical
Solutions, Mountain View, CA) was advanced into the right atrium to obtain
B-mode cine clips at the frame rate of 50-60 Hz were obtained. Images with four chamber views (Figure 19) were then offline analyzed using Syngo Vector Velocity
Imaging software (Siemens Medical Solutions, Mountain View, CA).
B-mode clips with adequate visualization of the endocardial border were studied.
Continuous echocardiographic monitoring ensured proper device fit and selection of qualified images. The endocardial LV borders were manually traced and verified at end-diastole to calculate Lagrangian strain and temporal derivative strain rate of the
LV. End-systolic and end-diastolic longitudinal strain values for all six segments of the LV (basal/mid/apical septal wall, and basal/mid/apical lateral wall) were recorded and averaged to obtain global longitudinal strains (GLS) for each trial.
59
Peak strain rates of global and regional LV wall, derived from the above strain data, were also analyzed in this study. LV opposing wall delay (difference in time-to-peak strain between mid septal and lateral walls, ms) was compared between experimental states during both systole and diastole.
LV Geometry Assessment
B-mode clips, collected by 2-D intracardiac echocardiography, allowed calculation of LV diameters and volumes. End-diastolic and end-systolic volumes, ejection fraction, and eccentricity (the ratio of long and short axis diameters of the LV chamber) were recorded and analyzed for each clip. Additionally, normalized volumes were derived based on body surface area and compared between states.
Statistical Analysis
JMP version 13 commercial software (SAS Institute, Cary, NC) was used for statistical analysis. All values are expressed as mean ± standard error of the mean
(SEM). Normality tests were performed prior to statistical comparisons. For normally distributed data, a two-sided, unpaired t test was used for comparing between two experimental states. Regional LV wall strains and peak strain rates were compared using one-way ANOVA with post hoc Tukey’s HSD tests. Differences with a two-sided p value of less than 0.05 was considered statistically significant.
60
RESULTS
A total of 12 canine subjects met selection criteria for this study. Studies where failure of ROSC after repeated defibrillation attempts were excluded. Basic pressure measures of two experimental states (Native beating heart versus DMVA support during VF arrest) are shown in Table 3. DMVA support of the fibrillating heart produced LV and arterial hemodynamics similar to the physiological state. There was no significant difference between two groups in terms of systolic and diastolic blood pressure, MAP, pulse pressure, LVESP, and LVEDP (all Ps > 0.05).
LV Geometry Characterization
We analyzed LV geometry (volumes and diameter of the chamber) of the native beating heart and DMVA supported VF arrested heart. LV volumes and normalized volumes (normalized to body surface area) were compared between two experimental states. SV is the difference of EDV and ESV. DMVA assist during VF consistently produced EDV, ESV, and SV similar to the physiological magnitude
(Figure 20A). The proportional relationship shown in the bar graph approximately reflected ejection fraction (EF=SV/EDV) values (49.8% for NBH group, and 45.6% for DMVA group). Volume index (normalized volume) is computed as the quotient of volume and BSA. EDVI, ESVI, and SVI showed no significant difference between DMVA and NBH group as expected (Figure 20B). Eccentricity index was compared at end-systole, which was calculated as long axis versus short axis
61
dimension of the left ventricle. Eccentricity index during DMVA support was found similar to the native beating heart (P=0.88) (Figure 20C).
LV Pump Function Assessment
LV pump function was compared between normal functioning heart and DMVA mechanically supported fibrillating heart (Table 4). DMVA assist generated SV, EF and CO similar to the physiological state (all Ps > 0.05). LV +dP/dtmax during systole and -dP/dtmax during diastole were both insignificantly different between
DMVA and NBH group.
LV Mechanics Analysis
Mean Value Comparison
Six segments of LV wall (basal septal, mid septal, apical septal, apical free, mid free, and basal free walls) were comprehensively studied. We found that DMVA support produced end-systolic longitudinal strain including all six regional (sRLS) and global strains (sGLS) and end-diastolic longitudinal strain including all six regional (dRLS) and global strains (dGLS) similar to those observed in the native beating heart. Consistently, peak longitudinal strain rate occurred during either systole or diastole was similar between DMVA support versus NBH, including all six segmental RLSRs and average GLSR. LV opposing wall delay, reflecting synchronous regional strains, was insignificantly different during either systole or diastole between NBH versus DMVA group.
62
Heat Map Comparison
LV all six segments in the apical long-axis plane are colored from green to red to represent higher absolute values of strain or strain rate to lower ones in the heat map. During systole, no significant differences were found among six regional strains in the group of either NBH or DMVA support, although DMVA assist produced more universal sRLSs throughout the LV wall (color difference were closer) while native beating heart showed more polarized sRLSs (color difference were wider) (Figure 21A and Figure suppl. 1). During diastole, dRLS of the apical free wall in the NBH was significantly higher than the mid septum (P=0.013). In addition, dRLS of the apical free wall during DMVA support of the fibrillating heart manifested significantly enhanced magnitude versus other five LV segments (Figure
21B and Figure suppl. 1).
Figure 22 illustrates distribution of peak strain rates of the NBH and DMVA supported fibrillating heart. During systole, there was no significant difference among all six LV segments in the native beating heart (Figure 22A). DMVA support of the fibrillating heart, on the other hand, generated peak strain rate of the apical free wall significantly higher than the mid septal and free walls (Figure suppl. 2).
During diastole, NBH exhibited insignificantly different regional peak strain rates among six LV regions. In contrast to the universal observed in the NBH, apical free wall of the DMVA supported fibrillating LV showed significantly elevated regional peak strain rate than mid free wall (P<0.001) (Figure suppl.2).
63
DISCUSSION
The aim of this study was to characterize left ventricular myocardial mechanics during mechanical ventricular actuation of the fibrillating heart. DMVA, having been shown to be the effective resuscitative support device, has advantages of directly actuating the ventricles and imposing wall motion to achieve life-sustaining hemodynamic support (through providing biventricular pump function). Analyzing
LV wall motion patterns during DMVA support is one means of determining the comparability of LV mechanical dynamics to the native beating heart. VF arrested hearts manifest akinetic properties and ventricular mechanics is entirely generated by external mechanical support. The results may further identify mechanisms whereby DMVA support may facilitate return of spontaneous circulation.
Although prognosis of either out-of-hospital or in-hospital CA patients remains poor, rapid resumption of adequate organ perfusion primarily determines neurologic and cardiac outcomes, and eventually, survival rate of CA and quality of life (Tomaselli
2015). In view of the urgent resuscitative demand and complexities of underlying pathophysiological conditions, an MCS device that allows prompt installation and restoration of circulation/perfusion would be optimal for resuscitative purpose. Such resuscitative devices should provide hemodynamic support without secondary myocardial injury. Additionally, such a device would ideally be associated with low
64
complication rates, given that these may outweigh the potential therapeutic benefits
(Ouweneel et al., 2012). Cardiac arrest treated with blood-contacting MCS devices such as ECMO, requires anticoagulation to avoid blood clotting, thrombosis, and thromboembolic complications. Blood contacting pumps also cause hemolysis and immune compromise as well as potential secondary life-threatening infections
(Nagpal et al., 2017). DMVA does not contact the blood, which has significant advantages. Utilizing similar actuation strategy of pneumatic artificial muscle applied in the mock heart simulation (Baturalp et al., 2015), DMVA directly encompasses the ventricles of the heart, compresses and expands them with an idealized systolic/diastolic forces, and provides near-physiological ventricular pump function as shown in this dissertation. Notably, commercially available acute MCS devices provide cardiac support by bypassing the circulation with blood pumps that contact the blood and predominant generated non-pulsatile flow. DMVA, on the other hand, mechanically actuates the heart, providing the mechanical work to the impaired or asystolic ventricles. Laboratory studies have previously indicated DMVA’s potentially beneficial effects for myocardial recovery, vital organ and cerebral resuscitation compared to other MCS devices and standard resuscitation techniques
(e.g., OC-CPR and CC-CPR).
This study utilized speckle tracking echocardiography (STE) interrogation to evaluate LV pump function and myocardial mechanics during DMVA support of the
VF arrested heart. Two-dimensional echocardiographic imaging provided an
65
objective means for characterizing changes in ventricular volumes while STE provided a means for measuring LV longitudinal strain. Results indicated that
DMVA support of the fibrillating heart generated similar long axis/short axis dimension ratio, EDV, ESV, and SV when compared to the native beating heart
(Figure 20). In other words, DMVA compressed and dilated the LV in a fashion that resulted in end-systolic and end-diastolic morphology similar to the native beating heart respectively. Furthermore, strain analysis revealed similarities between
DMVA supported hearts and the native beating heart with respect to all other measured strain parameters. There was no significant difference between NBH and
DMVA supported fibrillating hearts with respect to systolic and diastolic global and regional strain (Table 4). It is noteworthy that DMVA mechanically generated peak strain rates during support of the VF arrested heart similar to the NBH. These similarities were with respect to both systolic and diastolic, global and regional strain analyses (Table 4). These observations demonstrate that DMVA’s pneumatic forces can result in myocardial mechanics generated by the myocardium in the native beating heart.
Therefore, with appropriate delivery of drive system pneumatic forces DMVA can support the fibrillating heart and achieve LV pump function similar to that of the native beating heart. Additionally, in such circumstances, the resulting vascular hemodynamics generated by DMVA support during VF should be similar to the normal physiological state. In the studies presented in this dissertation, DMVA
66
support of the fibrillating heart generated LV and arterial hemodynamics (LVESP,
LVEDP, and systolic/diastolic blood pressures) similar to the physiological state
(Table 1). In addition, near-physiological EF, CO, LV +dP/dtmax, and LV -dP/dtmax were produced by DMVA assist compared to the normal functioning heart (Table
4). All pump function analyses during DMVA support of the fibrillating heart implies that the intrinsic properties of the myocardial tissues (inotropy or lusitropy) would be a direct result of DMVA’ effect on the myocardium. Therefore, references to DMVA’s effect on the myocardium during ventricular fibrillation will be termed mechanical inotropy for systolic actuation and termed mechanical lusitropy for diastolic actuation.
One expected standard for evaluation of LV contractility is the end-systolic pressure volume relationship (ESPVR). The slope of the relationship line defines the maximum elastance, which is a load-independent parameter of LV contractility. However, cardiac catheterization is an invasive procedure and requires immobilization, which limits the clinical application for assessing dynamic cardiac function. Notably,
ESPVR only describes the interaction of LV pressure and volume as pertaining to global LV pump function. This does not allow assessment of regional myocardial function of the heart. This becomes a disadvantage when addressing regional myocardial ischemia or other causes of regional dysfunction such as structural heart disease that do not pertain to the study. For the normal functioning heart with no other such underlying pathological circumstances, strain analyses are particularly ideal for assessing relatively load-independent global and regional pump function characteristics. Strain imaging has evolved and allows dynamic monitoring in both
67
laboratory and clinical settings using sophisticated intravascular probes and complex computer analyses. (Leung et al., 2010; Smiseth et al., 2016). Strain and its temporal derivative strain rate serve as objective functional indices of ventricular/myocardial function and dynamics. They clearly are superior to more elementary echocardiographic analyses such as ejection fraction and regional fractional shortening. Such more elementary indices have been utilized as indicators for risk stratification of patients with heart failure that is impacted largely by loading conditions. Measurement of GLS by 2-D STE has emerged as a sensitive index of myocardial systolic performance and to predict mortality and adverse cardiovascular outcomes in various patients populations (Cimino et al., 2013; Biering-Sørensen et al., 2017; Park et al., 2018). Strain profiles may also provide more perspectives regarding the regional wall motion relationships (such as mechanical synchrony).
Importantly, systolic strain rates provides a relatively load independent measures of myocardial function. This is explained by the fact that the final process of ventricular ejection occurs as a result of inertial effects of myocardial contraction. When myocardial contraction is complete, peak systolic strain rate which correlates with the rapid buildup of force in early systole occurs. This event is less influenced by afterload than strain or other inotropic parameters such as +dP/dtmax (Hamlin et al.,
2012). In our study, these measures were also used to interrogate the ventricular pump function and regional wall motion relationships imposed by DMVA support of the fibrillating heart. Thereby, this load-independent ventricular/myocardial functional analysis provided a logical means for comparing the DMVA supported fibrillating heart to the native beating heart.
68
Appropriately delivered DMVA forces to the fibrillating heart can, in fact, generate near-normal mechanical LV myocardial mechanics as quantitated in this study by systolic peak strain rate (sGLSR and sRLSRs for all six LV segments) and +dP/dtmax in this study (Table 4). Additionally, the fibrillating LV manifested near-physiological lusitropic properties (dGLSR, dRLSR, and -dP/dtmax) during DMVA support.
Intriguingly, DMVA did not alter the synchrony indices measured in this study.
Regional myocardial systolic and diastolic mechanics including LV opposing delay
(Table 4) during DMVA support also reflected that of the native beating heart.
This study demonstrated the unique effects of DMVA on myocardial mechanics and pump function, yet it has several limitations or unfinished tasks which ensure future investigations. First of all, comparisons using average values would be more indicative when integrated to waveform analyses. Specifically, for dynamic parameters such as flow, pressure, or deformation, introducing a comprehensive average waveform analysis will potentially reveal more information regarding
DMVA’s effects and device development. Aortic power (the product of simultaneous aortic flow and pressure) waveform analyses and average strain/strain rate waveform comparisons are warranted in future research. Secondly, the apical lateral region tended to show higher RLS and RLSR versus other LV segments during DMVA support, a very interesting regional target to interrogate in relation to the structure and actuating behavior of the DMVA cup. One explanation would be related to LV and RV interaction since DMVA is pumping both ventricles
69
spontaneously. Imaging tools to capture DMVA surface motion may offer more insights into the device’s effect on this specific LV segment.
Last but not least, cellular investigation is felt necessary to reveal the intertwined signaling pathways associated with mechanotransduction. Notably, although this study concentrates on DMVA’s effect on the fibrillating heart which has no effective pump function, the myocardium is still alive and capable of cellular responses such as mechanotransduction. Mechanotransduction refers to the molecular mechanisms by which cells response to stimuli in their physical environment by converting mechanical forces into biochemical or electrical signals.
Specifically, mechanical forces sensing from environmental factors, including osmotic pressure, tension, compression, and fluid shear stress, are major modulators of cellular function. Sensing these forces is essential to mediate cell proliferation and to maintain homeostasis in many tissues including cardiac muscle. Cyclic external stretch on a single normal cardiomyocyte during diastole triggered steady-state ROS production (microtubule integrity is essential), and a corresponding burst of Ca2+ sparks (increasing spark rate) from ryanodine receptors
(RyR2) located on the sarcoplasmic reticulum (Prosser et al., 2011, Prosser et al.,
2013). Mechanical forces may also acutely affect the function and conformation of sarcoplasmic/endoplasmic reticulum Ca2+-ATPase (SERCA2a), as well as the
I-band region of the giant protein titin in the sarcomere by either phosphorylation, redox modification, or SUMOylation (Fukuda et al., 2010; Sharov et al., 2006; Kho
70
et al., 2011). DMVA support has been demonstrated to provide both systolic compression and diastolic expansion to the whole LV wall that consists of living myocardial tissues. Therefore, further studying cellular effects of DMVA support of the fibrillating heart in relation to myocardial mechanics would help to better understand the underlying mechanism of myocardial protection and recovery post
ROSC.
CONCLUSION
In this study, DMVA support during VF arrest resulted in myocardial mechanics and regional LV wall motion relationships similar to the native beating heart.
Hemodynamics during DMVA support of the fibrillating heart were also restored to the physiological state. DMVA’s ability to restore near normal systolic and diastolic
LV pump function during VF arrest explains its unique capability for returning physiologic pulsatile perfusion during resuscitative circulatory support. Restoring relatively normal LV wall mechanics to the fibrillating heart during DMVA resuscitative circulatory support may benefit myocardial recovery following return of spontaneous circulation.
71
Figure 16. Schematic of the DMVA support system. The silicone constructed
DMVA cup provides both active systolic and diastolic support of the failing or fibrillating heart via delivery of atraumatic mechanical forces to the epicardial surface. Echocardiographic illustrations show canine end-diastolic and end-systolic four chamber views during DMVA support.
72
Figure 17. Experimental design. Two states were studied in this work: baseline
(blue, post-sternotomy preceding the initial fibrillation cycle) and DMVA support during fibrillation (green). Baseline data was collected of the native beating heart.
Animals then were subjected to 5-minute of ventricular fibrillation with total circulatory arrest followed by DMVA support. This design allows 3-5 arrest/recovery cycles to test DMVA during ventricular fibrillation (VF).
73
Figure 18. Experimental instrumentation. Animals were instrumented for aortic flow and pressure, and left ventricular (LV) pressure (not shown due to compressed
LV). An intravascular ultrasound catheter was placed down the left jugular vein and advanced into the right atrium to acquire intracardiac echocardiography.
74
Figure 19. Four chamber echocardiogram views of (A) native beating heart, (B) arrested unsupported fibrillating heart, and (C) DMVA supported VF arrest heart. LV: left ventricle; RV: right ventricle; VF: ventricular fibrillation. Note the similarities between the native beating heart and the fibrillating heart during DMVA support.
75
Table 3. Basic pressure characteristics of native beating heart and DMVA
support during VF. Values are presented as mean ± SEM. Comparisons used
unpaired t-tests. VF: ventricular fibrillation; LVESP: left ventricular end-systolic
pressure; LVEDP: left ventricular end-diastolic pressure. There was no significant
difference between two groups for each pressure measure.
Native Beating DMVA Support P Value Heart during VF arrest
Systolic Blood Pressure (mm Hg) 119.4 ± 5.6 118.1 ± 3.1 0.834
Diastolic Blood Pressure (mm Hg) 51.6 ± 5.4 46.7 ± 3.0 0.423
MAP (mm Hg) 74.2 ± 4.4 70.5 ± 2.4 0.453
PP (mm Hg) 67.8 ± 7.0 71.4 ± 3.8 0.653
LVESP (mmHg) 119.1 ± 6.2 118.0 ± 2.8 0.875
LVEDP (mmHg) -5.4 ± 3.2 -10.5 ± 1.4 0.066
76
Figure 20. LV geometry profiles of native beating heart and DMVA support during VF arrest. (A) End-systolic volume (ESV), end-diastolic volume (EDV), and stroke volume (SV) were compared between two groups. The entire bar represents EDV given that EDV is the sum of ESV and SV. The figure shows no significant difference between groups for all three parameters. (B) Normalized volume comparison was shown. Letter I means index. End-diastolic volume index
(EDVI) is the quotient of EDV and body surface area (BSA). There is no significant difference between groups for each parameter. (C) Eccentricity index is calculated as long axis versus short axis diameter. No significant difference was found at end-systole during DMVA compression.
77
Table 4. Left ventricular pump function and myocardial mechanics. Values are expressed as mean ± SEM. Comparisons used unpaired t-tests except for regional strain and strain rate which used one-way ANOVA and Tukey’s HSD tests. LV: left ventricle; VF: ventricular fibrillation; SW: septal wall; FW: free wall.
Native Beating DMVA Support P Heart during VF Value LV Pump Function Stroke Volume (ml) 15.5±0.9 14.5±0.8 0.440 Ejection Fraction (%) 49.8±1.9 45.6±1.6 0.100 Cardiac Output (L/min) 3.32±0.11 3.1±0.06 0.135
+dP/dtmax (mmHg/s) 1745.8±459.2 1709.5±697.0 0.971 -dP/dtmax (mmHg/s) -1665.6±672.6 -1771.5±590.2 0.739 LV Global Longitudinal Strain (%) End-Systolic -13.8±0.6 -12.4±0.5 0.080 End-Diastolic 10.1±0.9 11.8±0.5 0.124 LV Regional Longitudinal Strain (%) End-Systolic Basal-SW -15.9±1.3 -12.7±1.1 0.756 Mid-SW -12.3±1.4 -11.5±1.1 1.0 Apical-SW -10.4±1.4 -11.8±1.1 1.0 Apical-FW -14.8±1.3 -11.4±1.1 0.714 Mid-FW -15.4±1.3 -10.6±1.1 0.167 Basal-FW -10.3±1.4 -12.7±1.1 0.971 End-Diastolic Basal-SW 9.8±1.4 9.4±0.9 1.0 Mid-SW 4.2±1.4 7.8±1.1 0.604 Apical-SW 7.5±1.3 10±1 0.928 Apical-FW 11.4±1.3 15.9±1.1 0.219 Mid-FW 5.8±11.4 7.1±1 1.0 Basal-FW 10±1.3 8.9±1 1.0 LV Global Longitudinal Peak Strain Rate (1/s) Systolic -1.81±0.08 -1.98±0.07 0.119
78
Diastolic 1.82±0.07 1.96±0.06 0.128 LV Global Longitudinal Peak Strain Rate (1/s) Systolic Basal-SW -2.06±0.14 -1.82±0.11 0.960 Mid-SW -1.52±0.14 -1.76±0.11 0.971 Apical-SW -1.52±0.15 -1.92±0.10 0.568 Apical-FW -2.11±0.14 -2.33±0.11 0.981 Mid-FW -1.88±0.13 -1.58±0.10 0.880 Basal-FW -1.71±0.15 -2.00±0.09 0.920 Diastolic Basal-SW 1.93±0.14 1.88±0.11 1.0 Mid-SW 1.66±0.14 1.71±0.10 1.0 Apical-SW 1.77±0.15 2.10±0.11 0.802 Apical-FW 2.11±0.13 2.31±0.12 0.989 Mid-FW 1.84±0.14 1.58±0.10 0.956 Basal-FW 1.77±0.15 2.00±0.11 0.984 LV Opposing Wall Delay (ms) Systolic 144.8±16.2 158±12.2 0.209 Diastolic 144.7±16.5 177.7±13.5 0.124
79
Figure 21. LV strain heat maps: (A) regional end-systolic longitudinal strain
(sRLS, %); (B) regional end-diastolic longitudinal strain (dRLS, %); (C) color scale. Values are presented as mean. There were no significant differences for all six segments between NBH and DMVA during VF arrest.
80
Figure 22. LV strain rate heat maps: (A) regional peak longitudinal systolic strain rate (sRLSR, 1/s); (B) regional peak longitudinal diastolic strain rate
(dRLSR, 1/s); (C) color scale. Values are presented as mean. There were no significant differences for all six segments between NBH and DMVA during VF arrest.
81
Figure suppl 1. Regional wall strain intra-group comparisons. (A) Left ventricular (LV) longitudinal strain of the native beating heart during systole and diastole; (B) LV longitudinal strain of the DMVA supported VF arrested heart during systole and diastole. Asteroid sign indicates that two compared regions were significantly different. Cross sign indicates that the selected region was significantly different than all other regions. GLS: global longitudinal strain; SW: septal wall;
FW: free wall.
82
Figure suppl 2. Regional strain rate intra-group comparisons. (A) Left ventricular (LV) peak longitudinal strain rate of the native beating heart during systole and diastole; (B) LV peak longitudinal strain rate of the DMVA supported
VF arrested heart during systole and diastole. Asteroid sign indicates that there was a significant difference between two compared segments. GLSR: global longitudinal strain rate; SW: septal wall; FW: free wall.
83
CHAPTER V Left Ventricular Diastolic Function Is Returned during Direct Mechanical Ventricular Actuation of the Arrested Heart
This chapter addresses SA 1 of this dissertation, focusing on diastolic support.
Specific Aim 1: Determine if DMVA resuscitative support during Ventricular
Fibrillation (VF) results in left ventricular pump function similar to that of the native beating heart (NBH).
Aim1a – Determine if DMVA support of the fibrillating heart results in left ventricular diastolic filling similar to that of the NBH.
Aim1c – Determine if DMVA support of the fibrillating heart results in global and regional LV wall mechanics similar to that of the NBH.
Aim1d – Determine if DMVA support of the fibrillating heart results in LV inter-regional wall motion relationships similar to those of the NBH.
84
RATIONALE
Cardiac arrest (CA) is a devastating complication of cardiovascular disease, manifesting as the cessation of cardiac mechanical activity along with the confirmation of the absence of signs of circulation. In the United Sates, CA affects more than half a million populations annually with an extremely poor survival rates even after achieving return of spontaneous circulation (ROSC) (Benjamin et al., 2018;
Jentzer et al., 2018b). Incidence of out-of-hospital cardiac arrest (OHCA) reported by the US Resuscitation Outcomes Consortium sites indicated that 140.7 individuals per
100 000 population, or 347 322 adults, experience cardiac arrest (Mozaffarian D et al.,
2016). However, survival to hospital admission and hospital discharge are merely
29.0% and 10.8% respectively (CARES. 2017). Patients post successful resuscitation still experience multiple medical conundrums related to critical illness, including neurological deficits induced impairments of consciousness and cognition. Besides anoxic brain injury, post-arrest myocardial dysfunction (PAMD) has been discovered as another major functional impairment associated with poor survival rate and quality of life (Neumar et al., 2008). Post-resuscitation myocardial diastolic dysfunction was first demonstrated in porcine PAMD models (Kern et al., 1997; Xu et al., 2008) by utilizing echocardiography. Clinical studies have also suggested that diastolic dysfunction measured by Doppler-based echocardiographic may be associated with long-term mortality after OHCA (Chang et al., 2007; Jentzer et al., 2018a).
85
Direct mechanical ventricular actuation (DMVA) is a non-blood-contacting ventricular compression device that provides both systolic and diastolic resuscitative support
(Anstadt et al., 1995; Anstadt et al., 2008; Zhou et al., 2018b) (Figure 23). The device has been demonstrated to augment LV diastolic function of the failing heart (Anstadt et al., 2009; McConnell et al., 2014). Furthermore, DMVA clearly augments LV filling during support of the arrested or fibrillating heart. However, diastolic pump function of the fibrillating left ventricle during DMVA support has not been well characterized.
This study sought to determine if the application of DMVA to the fibrillating, arrested heart restores LV diastolic function to the native functioning heart in a clinically-relevant large animal model.
RESEARCH DEISGN
The Wright State University Lab Animal Care and Use Committee approved the experimental protocol, and all animals were treated in compliance with the Guide for the Care and Use of Laboratory Animals of the National Academies, published by the
National Institutes of Health, revised 2011. All surgeries were performed by a qualified, trained, and experienced cardiothoracic surgeon team.
VF Cardiac Arrest Model
Adult Yorkshire swine (n=8, 80±2kg) were selected as they provide heart sizes similar to the average adult human. Animals were pre-anesthetized with Telazol 6
86
mg/kg, Xylazine 3 mg/kg, and Atropine 0.02 mg/kg. Anesthesia during the experiment was maintained by 1-2% isoflurane (Chapter III, surgical protocol).
Subjects then underwent median sternotomy, pericardiotomy, and instrumentation for hemodynamic and intracardiac echocardiography (Figure 24). A TS420 perivascular flowmeter (Transonic Systems Inc, Ithaca, NY) connected to an ultrasound probe was placed on the ascending aorta. P23XL transducers
(Spectramed Inc, Mt Vernon, OH) were introduced into both the aorta and LV via the right carotid. A 10-MHz Acuson AcuNav10F intravascular ultrasound catheter
(Siemens Healthcare) was advanced into the right atrium. Following baseline recordings of the native beating heart (NBH), VF arrest was induced using a 9-volt electrical shock for 5 min of circulatory arrest. DMVA was then applied for 15 min during VF arrest followed by defibrillation attempts. After return of spontaneous circulation (ROSC), hemodynamics was monitored for 15 minutes. VF arrest was then resumed and repeat the above cycle. The entire resuscitation trail typically took three to four hours allowed for multiple repeats of data collection. Animals were euthanized by potassium chloride injection under anesthesia at the end of the experiment (Beaver et al., 2001). DMVA support rates were adjusted to 80 times per minute, which is consistent with natural rhythms of swine. Swine subjects used
105-100 mm diameter cups. A low intensity apical vacuum generated a tight interface between the heart and DMVA inner membrane.
87
2-D Intracardiac Echocardiography
Intracardiac echocardiography including speckle tracking imaging was interrogated on the swine CA model utilizing an ACUSON Sequoia C512 System (Siemens
Medical Solutions, Mountain View, CA). B-mode cine clips (50-60 Hz) with four chamber view (Figure 25) were obtained and processed offline utilizing Syngo
Vector Velocity Imaging software (Siemens Medical Solutions, Mountain View,
CA). B-mode clips with adequate resolution of the endocardial border were studied.
The endocardial LV borders were traced for multiple times at end-diastole to ensure accurate acquisition of myocardial deformation.
Diastology Analysis
Data of hemodyncamics, LV geometry, and LV mechanics were collected, computed, and compiled using different modalities. LV diastolic function was determined by analyzing LV profiles of end-diastolic volume (EDV), end-diastolic volume index (EDVI), end-diastolic pressure, -dP/dtmax, end-diastolic global longitudinal strain (dGLS), peak global longitudinal strain rate (dGLSR), end-diastolic regional longitudinal strain (dRLS), peak global longitudinal strain rate (dRLSR), and diastolic dyssynchrony index (the standard deviation of delay in time to peak strain between basal septum and all other regions).
88
Hemodynamic Data Collection
Pressure and flow data were periodically collected throughout the experiments for
10 sec intervals at 200 Hz using LabVIEW data acquisition software (National
Instruments, Austin, TX). Waveform derived parameters including cardiac output
(CO), mean arterial pressure (MAP), LV end-diastolic pressure (LVEDP), and LV
-dP/dtmax were analyzed to compare between DMVA and NBH.
Diastolic LV Geometry Assessment
B-mode clips, collected by 2-D intracardiac echocardiography, allowed calculation of LV diameters and volumes. EDV and normalized EDV (devided by body surface area) were computed to compared DMVA versus NBH.
Diastolic Myocardial Mechanics Assessment
Speckle tracking echocardiography (STE) was utilized to calculate Lagrangian strain and temporal derivative strain rate of the LV. End-diastolic longitudinal strain values for all six segments of the LV (basal/mid/apical septal wall, and basal/mid/apical lateral wall) were recorded and averaged to obtain global longitudinal strain for each trial. Peak strain rates of global and regional LV wall, derived from the above strain data, were also analyzed in this study. Dyssynchrony index, reflecting LV regional wall motion relationship, was compared between experimental states during diastole.
89
Statistical Analysis
JMP version 13 commercial software (SAS Institute, Cary, NC) was used for statistical analysis. All values are presented as mean ± standard error of the mean
(SEM). Normality tests were taken prior to statistical comparisons. For normally distributed data, a two-sided, unpaired t test was used for comparing between NBH and DMVA. Regional LV strains and peak strain rates were compared using one-way
ANOVA with post hoc Tukey’s HSD tests. For all analyses, significance was set at a two-sided p value of less than 0.05.
RESULTS
A total of 8 animals met selection criteria for this study. Two porcine subjects were excluded due to major bleeding or pericardial laceration (a consideration when working with the species). In general, porcine post-sternotomy baseline was in a hypotensive condition with relatively normal cardiac output. Hemodynamics of two experimental states (native beating heart versus DMVA support during CA) are shown in Table 5. DMVA support generated approximately 78.6% baseline cardiac output, although it was significantly lower than the native beating heart. Arterial pressures including systolic, diastolic, and mean arterial pressure during DMVA support of the arrested heart were significantly reduced compared to NBH group.
There was no significant difference in pulse pressure between two states.
90
LV Geometry Assessment
LV volumes and normalized volumes (normalized to body surface area) were calculated and compared between two groups. SV is computed as the difference of
EDV and ESV. EDV, ESV, and SV consistently showed no significant difference between NBH and DMVA support (Figure 26A). Volume index is the quotient of volume and BSA. As expected, DMVA support of the fibrillating heart provided
EDVI, ESVI, and SVI similar to the physiological state (Figure 26B). Ejection fraction (EF=SV/EDV) could be computed and clearly identified from the proportional relationship shown in the bar graph, which were 46.2% and 48.3% for
NBH and DMVA group respectively.
Diastolic Function Evaluation
LV diastolic function was compared between native functioning heart and DMVA mechanically supported arrested heart (Table 6). DMVA support generated LVEDP similar to the physiological state (P=0.529). The maximum slope of LV pressure during diastole (-dP/dtmax) was insignificantly different between two groups
(P=0.233). LV diastolic dimensions including EDV, EDVI, and end-diastolic long/short axis diameter ratio were all similar during DMVA support when compared to NBH. Myocardial mechanic profiles were analyzed in both groups to reveal more functional similarities or discrepancies. We found that DMVA support produced end-diastolic longitudinal strain including six regional (dRLS) and global strains (dGLS) similar to those observed in the native beating heart. Importantly,
91
peak longitudinal strain rate was significantly increased during DMVA support versus NBH including all six segmental dRLSRs. As a result, DMVA support during VF arrest manifested more enhanced dGLSR than the native beating heart. It is noteworthy that synchronous relationship among diastolic regional peak strain
(dyssynchrony index) was not significantly elevated during DMVA assist when compared to NBH (P=0.071).
Spatial Distribution of LV Myocardial Mechanics
LV heat maps were utilized to display dRLS and dRLSR of all six segmental LV wall (basal lateral, mid lateral, apical lateral, basal septal, mid septal, and apical septal wall). LV segments are colored from green to red to represent higher magnitudes of strain or strain rate to lower ones. In general, lateral wall tended to show higher dRLS versus septum, which was observed in both native beating heart and DMVA support during VF arrest (Figure 27A). Statistically, no significant differences were found among six regional strains in the state of either NBH or
DMVA support (Figure 28), although one reverse tendency was observed in the basal septum (higher magnitude than other regions) of the native beating heart.
DMVA assist globally and regionally generated more robust strain rate than that of the NBH, which was clearly shown by the color difference (DMVA supported heart displaying all areas as green, while NBH as red and yellow) (Figure 27B). For intra-group comparison of regional peak strain rates, dRLSR in apical lateral wall of
92
the native beating heart was significantly higher than mid or basal lateral wall
(Figure 29A), while DMVA supported arrested heart manifested significantly elevated apical septal dRLSR versus either mid or basal septal dRLSR (Figure 29B).
Notably, all dRLSRs in either NBH or DMVA group were shown insignificant difference when compared to each dGLSR respectively.
DISCUSSION
This study demonstrated the capability of DMVA support of the fibrillating heart to restore diastolic function of the fibrillating porcine heart. DMVA effectively provides diastolic actuation to the arrested heart and augments LV diastolic myocardial mechanics, observed in a previous canine CA model. We continued to analyze DMVA resuscitative effects on the porcine heart, which shares the similar size and anatomical structure of human adults. The experimental animals were supported by DMVA with a simulated systolic/diastolic duration ratio close to the physiological level. During mechanical diastole generated by external DMVA support, the arrested heart achieved 78.6% baseline cardiac output. Although
DMVA resulted in relatively hypotensive state compared to baseline (prior to VF arrest), it generated arterial pulsatility (i.e. pulse pressure and surplus hemodynamic energy) similar to the native beating heart at baseline (Table 5). Surplus hemodynamic energy (SHE) quantitates the additional energy that maintains pulsatile flow to ensure more physiologic microcirculation and better vital
93
end-organ recovery than continuous flow. Of note, hemodynamics including systemic pressure and aortic flow during DMVA support of the arrested canine heart was returned to the physiological state (data not shown). It is intriguing that DMVA support returned blood pressure only below the baseline level, not surpassing that.
This observation was possibly attributed to the significantly lower cardiac output achieved by DMVA support, or impaired sympathetically-mediated vasoconstriction post CA (either sympathetic nerve activity or vascular responsiveness, or both).
Diastology represents the process of LV filling during diastole, which includes four phases: isovolumic relaxation (active relaxation that requires energy), rapid filling, diastatsis, and atrial systole. LV diastolic function was first quantitated by LV filling pressure in this study. LV filling pressure is mainly affected by LV chamber stiffness/compliance. In the acute CA model, the fibrillating heart did not occur myocardial hypertrophy. LV filling pressure, therefore, was directly produced by external mechanical actuation. Specifically, DMVA support mechanically generated suction forces without restricting the LV dimension during diastole and thus resulted in LVEDP similar to the native beating heart (Table 6). LVEDP has been found to be the only abnormally elevated pressure because of a large atrial pressure wave, while other LV filling pressure indicators such as mean pulmonary capillary wedge pressure, left atrial pressure, and LV pre-A pressure remain normal (Nagueh et al.,
2016; Andersen et al., 2017). Normal LVEDP achieved by DMVA support, along
94
with physiological -dP/dtmax (Table 6), implies that the mechanical circulatory support returned LV diastolic filling when actuating the fibrillating heart.
LV chamber characteristics, being another determinant of diastology, were compared between the native beating heart and DMVA actuated VF arrested heart.
Similar EDV, EDVI, and end-diastolic long axis/short axis dimension ratio were achieved during DMVA support versus baseline, which is consistent with the fact
DMVA cup fully expanded the LV without any restriction at the end of mechanical diastole. Volume changes in relation to LV chamber compliance also explained physiological LVEDP generated by DMVA actuation. Ejection fraction represents
LV volume changes during either native contraction or mechanical compression of the heart. Notably, EF being 48.3% in DMVA group (46.2% in NBH group) was considered at the normal range for swine.
Myocardial mechanics were comprehensively analyzed in this study by utilizing
2-D STE. STE provides both static and dynamic measures of diastolic function.
End-diastolic strain characterized ventricular loading achieved prior to ventricular contraction (NBH) vs compression (DMVA during VF arrest). Global and regional peak diastolic strain rates offered dynamic measures of lusitropy. Note that all diastology properties during DMVA support were mechanically achieved, and thus any intrinsic properties of the myocardial tissues (i.e. lusitropy) would be referred as mechanical lusitropy when discussing DMVA’s effects. Strain analysis discovered
95
consistent findings to volume changes given that strain is defined as the deformation of myocardial tissues normalized to its original shape and size. End-diastolic global and regional longitudinal strain for all six LV segments were similar between the native and mechanically supported LV. Furthermore, dRLS among six segments showed no significant difference in both groups (Figure 27 and 28). Asynchronous
LV wall motion and deformation can potentially be a concern for external mechanical support device. Mechanical myocardial synchrony served as a means for assessing regional myocardial interactions during diastole. This study revealed that
DMVA support did not increase diastolic asynchronous LV wall strain quantitated by dyssnchrony index (Table 6). Post-resuscitation myocardial diastolic dysfunction has been shown to be associated with both short-term and long-term mortality in patients resuscitated from cardiac arrest (Jentzer et al., 2015; Jentzer et al., 2018a).
DMVA support of the fibrillating heart resulted in LV diastolic dynamics which mimicked that of the normal beating heart. These effects may be advantageous for improving myocardial diastolic dysfunction post resuscitation following return of spontaneous circulation (ROSC).
LV diastolic longitudinal strain rate peaks during the isovolumic relaxation period
(early diastole). This phase of diastole involves active myocardial relaxation, which is an energy dependent process. Calcium is the key trigger leading to the active processes of both myocardial contraction and relaxation. During diastole, cytoplasmic calcium is pumped into sarcoplasmic reticulum within the myocyte by
96
sarcoplasmic/endoplasmic reticulum calcium ATPase (SERCA2a) or pumped out of the myocyte by the sarcolemmal sodium-calcium exchanger. These processes serve as an initiating signal to cease troponin C and other downstream activity, which entail that calcium is first dissociated from troponin C, Actin then detaches from myosin, and diastole is initiated. The main ATP consuming process during diastole is the sequestration of cytoplasmic calcium, which is largely dependent on the activity of the ATPase ion pump ---- SERCA2a, removing approximately 75% calcium in healthy human (Morrissey et al., 2016). Impaired calcium removal leads to diastolic dysfunction, manifesting as reduced dGLSR (Ersbøll et al., 2014). There are also elastic recoiling forces of I-band region of the giant protein titin that aid in relaxation (Iribe et al., 2007). Note that diastole does not consume ATP in relation to actin-myosin ATPase involved in the excitation-contraction coupling. The actin/myosin detachment only requires ATPs binding to myosin to take apart the crossbridge structure (ATPs stay binding to myosin to get ready for next cycle of systole). Importantly, DMVA support of the fibrillating heart causes the stretch of the living myocardium and ventricle during mechanical diastolic actuation. This mechanical effect induced by DMVA support may secondarily affect calcium sequestration through redox modification of the SERCA2a or phosphorylation of N2 domain of the titin I-band via acute mechanotransduction (Prosser et al., 2014;
Krüger et al., 2009). This cyclic mechanical stretch induced by DMVA support of the fibrillating heart to the underlying stunned myocardial tissues could attenuate myocardial diastolic dysfunction post resuscitation.
97
LV longitudinal peak diastolic strain rate derived from STE provides more easily derived interrogation of diastolic function than early diastolic mitral annular tissue velocity (e’) derived from tissue Doppler imaging (TDI) (Flachskampf et al., 2015;
Choudhury et al., 2017). The advantage of diastolic strain derived by STE is that the regional early rate of diastolic deformation is more easily and accurately derived.
Thereby, diastolic performance of the entire ventricle and all pertinent LV segments can be assessed. Specifically, STE avoids the limitations of angle dependency
(results impacted by cardiac translational motion) and inferior signal-to-noise ratio found in TDI (Mor-Avi et al., 2011). LV dGLSR measurements have been shown a significant association with the time constant of LV relaxation (τ). Additionally, ratio of peak transmitral early diastolic velocity (E) and dGLSR contributes important information about global myocardial relaxation (Wang et al., 2007). In this study, DMVA support of the fibrillating resulted in diastolic global and regional strain rates that exceeded that of the native beating heart. This effect can only be attributed to DMVA forces which resulted in enhanced myocardial lusitropy. LV chamber compliance could also be different during DMVA support of the fibrillating heart which may have contributed to these findings when compared to the native beating heart. Additionally, LV filling as indicated by -dP/dtmax during
DMVA support was similar to that of the native beating heart (Table 6), which is consistent with the added mechanical lusitropy observed using strain analyses.
98
Taken together, this study demonstrated that DMVA support providing mechanical diastole mimics the relaxation and the viscoelastic recoil properties of the native heart. In this way, DMVA could potentially return LV diastolic function (optimize preload) for adequate vital organ perfusion during resuscitation and prevent severe myocardial diastolic dysfunction post resuscitation. However, there are some limitations we want to address. First, DMVA support generated a partially different pattern of dRLSR (LV apical septal region higher than other two septal regions) on the akinetic heart when compared to the native beating heart (LV apical lateral region higher than other two lateral regions). One explanation would be related to the contralateral effect of RV during DMVA bi-ventricular support. Imaging modalities to monitor DMVA motion may shed lights on the specific LV segment that is affected by DMVA actuation. Secondly, investigations of molecular signaling pathways (targets such as SERCA2a and titin I-band) involved in the mechanotransduction are warranted to better understand the underlying mechanism of myocardial recovery post ROSC in relation to DMVA imposed myocardial mechanics. Finally, no radial or circumferential strain data was collected mainly due to the ultrasound interference of DMVA support shown in the B-mode clips. High frame rate ultrasound probe might be the solution in future research. Short-axis view of the echocardiogram offers informative data regarding LV wall torsional deformation which attributes to the suction effect of cardiac untwisting during early diastole.
99
CONCLUSION
This study found LV end-diastolic volume and strain were returned to the physiological range during DMVA support of the VF arrested heart. DMVA assist of the fibrillating LV also generated LVEDP and LV -dP/dtmax similar to the native beating heart. Most notably, diastolic peak strain rates were significantly enhanced during DMVA support vs the NBH. Furthermore, DMVA resulted in near-physiological patterns of diastolic mechanical synchrony. These findings may have important implications on recovery of diastolic function following DMVA resuscitative circulatory support and beneficial outcomes against post-arrest myocardial dysfunction.
100
Figure 23. Schematic of the DMVA support system. The DMVA support system consists of the silicone constructed DMVA cup and the attached drive system, which provides systolic and diastolic pneumatic forces that inflate and deflate the inner membrane of the DMVA cup, respectively. Echocardiographic illustrations show swine end-diastolic and end-systolic four chamber views during DMVA support.
101
Figure 24. Experimental instrumentation. Animals were instrumented for aortic flow and pressure, and LV pressure (not shown). An intravascular ultrasound catheter was advanced into the right atrium to obtain echocardiographic clips with four chamber view.
102
Figure 25. Four chamber view Echocardiogram Images of (A) native beating heart (end-diastole), (B) unsupported VF arrested heart, and (C) DMVA supported VF arrested heart (end-diastole) in swine. Note the similarities between the native beating heart and the arrested heart during DMVA support.
103
Table 5. Hemodynamic characteristics of native beating and DMVA supported
arrested hearts. Values are presented as mean ± SEM. Comparisons used unpaired
t-tests. VF: ventricular fibrillation; CO: cardiac output; MAP: mean arterial
pressure; PP: pulse pressure. SHE: surplus hemodynamic energy. * * p < 0.01, vs.
(퐹푙표푤×푃푟푒푠푠푢푟푒)푑푡 NBH. Equation: SHE = 1,332(∫ − MAP). ∫ 퐹푙표푤 푑푡
Native Beating DMVA Support P Value Heart during VF arrest
CO (L/min) 3.08 ± 0.14 2.42 ± 0.08** <0.001 Systolic Pressure (mm Hg) 96.5 ± 4.0 73.4 ± 1.4** <0.001 Diastolic Pressure (mm Hg) 59.3 ± 1.9 36.5 ± 0.7** <0.001 MAP (mm Hg) 71.7 ± 1.9 48.8 ± 0.7** <0.001 PP (mm Hg) 37.2 ± 4.3 36.9 ± 1.5 0.951 SHE (ergs/cm3) 20,651.9 ±2,768.2 15,156.5 ±1,531.6 0.075
104
Figure 26. LV geometry of native beating heart and DMVA support during
VF arrest. (A) End-systolic volume (ESV), end-diastolic volume (EDV), and stroke volume (SV) were compared between two groups. The entire bar represents EDV given that EDV is the sum of ESV and SV. The figure shows no significant difference between two experimental states for all three parameters. (B) Normalized volume comparison is shown. Letter I means index. End-diastolic volume index
(EDVI) is the quotient of EDV and the body surface area (BSA). There is no significant difference between groups for each parameter.
105
Table 6. LV diastolic function of native beating and DMVA supported arrested hearts. Values are presented as mean ± SEM. Comparisons used unpaired t-tests except for regional mechanics which used one-way ANOVA and Tukey’s HSD tests. SW: septal wall; LW: lateral wall. EDVI: EDV normalized to body surface area. * p < 0.05, vs. NBH; * * p < 0.01, vs. NBH.
Native Beating DMVA Support P Heart during VF Arrest Value End-Diastolic Pressure (mmHg) 2.43±0.95 4.34±0.75 0.529
-dP/dtmax (mmHg/s) -1371.7±653.9 -1674.4±490.1 0.233 End-Diastolic Volume (ml) 63.1±2.3 66.6±2.1 0.257 EDVI (ml/m2) 40.4±1.4 42.7±1.3 0.243 End-Diastolic Long/Short Axis Dimension Ratio 1.75±0.07 1.80±0.06 0.611 Global Longitudinal Strain (%) End-Diastolic 11.0±1.1 13.5±0.8 0.063 Regional Longitudinal Strain (%) End-Diastolic Basal-SW 14.4±1.8 11.1±1.3 0.950 Mid-SW 8.2±2.0 12.0±1.2 0.887 Apical-SW 10.2±2.2 11.7±1.3 0.887 Apical-LW 11.9±1.7 15.0±1.3 1.000 Mid-LW 10.0±2.1 13.2±1.3 0.978 Basal-LW 9.3±2.0 13.2±1.3 0.911 Global Longitudinal Peak Strain Rate (s-1) Diastolic 1.35±0.08 2.26±0.07** <0.001 Regional Longitudinal Peak Strain Rate (s-1) Diastolic Basal-SW 1.37±0.13 1.97±0.11* 0.024 Mid-SW 1.31±0.13 1.91±0.1* 0.018 Apical-SW 1.35±0.12 2.55±0.11** <0.001 Apical-LW 1.72±0.14 2.28±0.12* 0.044
106
Mid-LW 0.90±0.13 2.10±0.12** <0.001 Basal-LW 0.99±0.12 2.12±0.11** <0.001 Dyssynchrony Index (ms) Diastolic 182.7±10.5 196.9±10.7 0.071
107
Figure 27. LV heat maps: (A) regional end-diastolic longitudinal strain (dRLS,
%); (B) regional peak longitudinal diastolic strain rate (dRLSR, 1/s). Values are presented as mean. There were no significant differences for all six segmental strains between NBH and DMVA during VF arrest.
108
Figure 28. Regional diastolic strain intra-group comparisons. GLS: global longitudinal strain; SW: septal wall; LW: lateral wall. There was no significant difference among six LV regions in either NBH or DMVA support group.
109
Figure 29. Regional peak strain rate intra-group comparisons. Asteroid sign indicates that two compared regions has significant difference. GLSR: global longitudinal peak strain rate; SW: septal wall; LW: lateral wall. In native beating heart group (Left panel), apical-lateral wall showed significantly high dRLSR than mid-lateral and basal-lateral walls. In DMVA supported arrested heart group (Right panel), apical-septum showed significantly high dRLSR than mid-septal and basal-septal walls. Every regional peak strain was statistically similar to GLSR, found in both NBH and DMVA groups. Apical regions tend to display higher peak strain rate during diastole in NBH group, which is similar to what we discovered in
DMVA group.
110
CHAPTER VI Direct Mechanical Ventricular Actuation during Cardiac Arrest Generates Pulsatile Hemodynamics Similar to the Native Beating Heart
This chapter addresses SA 2 of this dissertation.
Specific Aim 2: Determine if DMVA resuscitative support of the fibrillating heart results in pulsatile hemodynamics similar to the physiological state.
Aim 2a – Determine if DMVA support of the fibrillating heart generates near-physiological pulsatile hemodynamics.
Aim 2b – Determine if DMVA support of the fibrillating heart results in pulsatile hemodynamics similar to return of spontaneous circulation (ROSC) at various blood flow dynamic states.
111
RATIONALE
With the advent of fully-implantable continuous flow ventricular assist devices
(VAD), the impact of pulsatility on human health garnered increased scrutiny.
Researchers are divided as to whether non-pulsatile flow is a significant contributor to common pathologies seen in continuous VAD use (Cheng et al., 2014; Kato et al., 2011; Soucy et al., 2013; Baric, 2014). However, comparatively little attention has been paid to the importance of acute pulsatility. Currently available devices for treating cardiac arrest (CA) and/or unexpected circulatory collapse utilize modified cardiopulmonary bypass (CPB) circuits that can be implemented using percutaneous catheter techniques. These devices have been used successfully for temporary resuscitative support or bridging to alternative therapies. However, neurologic injury and end-organ dysfunction remain a significant cause of morbidity and mortality in these scenarios. These complications are multifactorial but related in part to delay in returning effective organ perfusion.
Direct mechanical ventricular actuation (DMVA) is a non-blood contacting support device with unique efficacy for providing resuscitative circulatory support.
Installation is relatively simple and requires no anticoagulants (Anstadt et al.,
1991a; Anstadt et al., 1991b). Once applied, DMVA can rapidly restore normal hemodynamics. Prior investigations provide compelling evidence that DMVA improves cerebral resuscitation outcomes and neurological recovery compared to
112
cardiopulmonary bypass in animal models (Anstadt et al., 1990, Anstadt et al.,
1991c, Anstadt et al., 1992, Anstadt et al., 1993). These findings are felt secondary to DMVA’s ability to return physiologically normal pulsatile flow during resuscitation. The purpose of this study was to determine if DMVA resuscitative support to the fibrillating heart generates physiologic pulsatile flow similar to the native beating heart in a mature, clinically-relevant animal model.
RESEARCH DESIGN
The Wright State University Lab Animal Care and Use Committee approved the experimental protocol, and all animals were treated in compliance with the Guide for the Care and Use of Laboratory Animals of the National Academies, published by the
National Institutes of Health, revised 2011. All surgeries were performed by a qualified, trained, and experienced cardiothoracic surgeon team.
Resuscitation Model
Mature mixed-breed canine (n=15, 32±4.3kg) and Yorkshire swine (n=9, 85±2.5kg) were studied using a model of repeated periods of cardiac arrest (Figure 30). These species were selected to study cup sizes relevant to the expected range of the adult human population. Animals were anesthetized (1-2% isoflurane), underwent median sternotomy and pericardiotomy, and instrumented for hemodynamic monitoring
113
(Chapter III, surgical protocol). After collecting baseline data, ventricular fibrillation (VF) was induced using a 9 V shock for 5 mins of unsupported circulatory arrest. DMVA was then applied for a 15 min followed by defibrillation attempts. After return of spontaneous circulation (ROSC), hemodynamics were monitored for 15 mins. Fibrillation refractory to three successive countershocks was treated with another 15 minutes of DMVA support, after which defibrillation attempts were repeated. Other than the administration of standard doses of lidocaine, no other cardiovascular agents such as vasopressors or inotropes were administered during the experimental protocol. Intravenous fluids were administered to maintain central venous pressures in the physiologic range. This resuscitation cycle was repeated for 4 hours after which animals were euthanized by potassium chloride injection under anesthesia (Beaver et al., 2001).
Proprietary drive algorithms were used to control cup dynamics. DMVA rates were adjusted to match natural rhythms in animal subjects. Canine subjects had heart rates approximately 120 bpm while swine had rates approximately 80 bpm. Drive systolic duration remained at approximately 50% throughout the course of the experiments. Systolic actuation air pressures were adjusted to produce baseline systolic blood pressures. Support during fibrillation generally necessitates higher air pressures without the additive support provided when supporting the failing heart.
Diastolic air pressures were adjusted to fully retract the inner membrane to the semi-rigid outer shell to facilitate diastolic assist. Continuous intracardiac
114
echocardiography was used assess for proper device fit and monitor device mechanics throughout the experiment. Canines used 85-90 mm diameter cups
(measured at the rim of the cup near the atrioventricular groove) while Swine used
105-100 mm diameter cups. A low intensity apical vacuum suctioned the heart in the cup and generated a tight interface between the epicardial surface and device inner membrane.
Data Collection and Analysis
A TS420 perivascular flowmeter (Transonic Systems Inc, Ithaca, NY) was connected to an ultrasound probe placed on the ascending aorta to measure cardiac output (CO). A P23XL transducer (Spectramed Inc, Mt Vernon, OH) was introduced into the aorta via the right carotid. Aortic pressures and flows were periodically collected throughout the experiments for 10 sec intervals at 200 Hz using LabVIEW data acquisition software (National Instruments, Austin, TX). Each animal required at least one successful defibrillation attempt to be included in the final analysis. Data was grouped into three conditions: baseline, DMVA support during VF, and ROSC. DMVA and ROSC data were then further stratified into three flow groups: >75%, 50-75%, and 25-50% average baseline cardiac output for each animal. Inclusion criteria required steady-state conditions to be present for each 10 sec capture. Studies where failure of ROSC after repeated defibrillation
115
attempts were encountered, did not meet inclusion criteria and were therefore not included in the analysis.
Pulsatility Parameters
Grouped hemodynamic data underwent computational analysis for calculations of pulse pressure (PP), energy equivalent pressure (EEP), surplus hemodynamic energy
(SHE), pulsatility index (PI) and pulse power index (PPI) as defined in Table 7. For parameters using both pressure and flow, peak flows and pressures were identified using an automatic delineation algorithm (Kaebnick et al., 2007). Flows and pressures were aligned using cross-correlation to correct for phase shifts and standardize comparisons. Values for each state are displayed as mean ± standard error of the mean (SEM). A one-way ANOVA was run for each flow condition and post-hoc Tukey HSD tests were used to identify p values. To reject the null hypothesis, a two-sided p value of less than 0.05 (the significant level) was required.
Waveform Analysis
The described measures estimate pulsatility for an entire cycle. However, instantaneous changes throughout the cycle are masked by these measurements. The aligned aortic pressures and flows captures from the previous analysis were multiplied to produce aortic power. Waveforms were broken into individual cycles
116
using systolic peak location. Once isolated, aortic power cycles were standardized to a representative 700 ms cycle (to better capture information during interpolation) and resampled to 200 Hz to enable point-by-point comparisons. An ensemble average of the standardized aortic power waveforms produced an overall waveform defining that state (with standard error bars to represent signal variation).
RESULTS
A total of 16 animals (11 canine and 5 swine) met selection criteria for this study. A segment of the canine and swine subjects was excluded due to a lack of ROSC after repeated defibrillation attempts (a consideration when working with the species).
Since 24 animals in total underwent cardiac arrest, the success rate of ROSC post
DMVA support is promisingly 66.7% (73.3% in canine subjects and 55.6% in swine subjects respectively, see Table suppl. 1). Mean hemodynamics and pulsatility parameters were calculated during baseline, DMVA support during VF, and ROSC post support. Comparisons were made between baseline and each of the three flow categories during DMVA support and ROSC (Table 8). In the near-normal flow category, average aortic flow was significantly lowered (approximately 80.7% baseline flow, see Figure 31A) during DMVA support compared to beating heart equivalents (P<0.001), while mean arterial pressure (MAP) was insignificantly different between DMVA and baseline (P=0.15). Mean flows of DMVA supported animals during moderate and severe shock were significantly depressed versus
117
baseline (both Ps <0.001), with a respective proportion of 57.0% and 38.0% of the baseline. No significantly differences of aortic flow were observed between DMVA and ROSC in both reduced flow groups. It is noteworthy that differences in mean flows between DMVA and ROSC hearts became less pronounced in the moderate and severe shock hemodynamic groups.
Importantly, DMVA showed similar PP when compared to baseline in normal, moderate and low flow groups (P=0.15, 0.45 and 0.88, respectively), while significantly higher PP versus ROSC in all three flow groups (all Ps<0.001). We observed no statistically significant difference in EEP between DMVA support
(with near-normal flows) and baseline (P=0.26). Furthermore, EEP did not differ between DMVA and ROSC in any flow condition. SHE during DMVA was significantly higher than ROSC in every flow group (all Ps <0.001). As expected,
SHE of the supported animal did not differ from baseline in the near-normal flow condition (P=0.91). In moderate and severe depressed flow groups, however,
DMVA produced higher surplus energy than baseline did (P=0.002 and P<0.001, respectively). In the frequency domain analysis, DMVA generates similar PI as baseline when providing the near-normal flow (P=0.08). Interestingly, PIs of
DMVA supported animals were significantly increased compared to baseline in both moderate and severe shock categories (P=0.01 and P<0.001, respectively). In addition, PI showed significant differences between DMVA and ROSC in all three flow conditions. The findings of PPI, as a modification of PI with multiplied
118
corresponding angular frequency, were generally consistent with PI among three states, except that there was no significant difference between DMVA and ROSC in the near-normal flow category (P=0.11).
Average power curves were calculated for baseline and for each flow condition during DMVA support and ROSC (as displayed in Figure 32). Aortic power waveforms in the near-normal flow condition indicates that DMVA produced a much steeper downslope in late systole versus baseline and ROSC, while ROSC had a significantly blunted peak region compared to either baseline or DMVA support.
Power declined substantially from baseline in both moderate and severe shock.
However, DMVA has noticeably higher peaks than ROSC except during severely-depressed flow. While the early half of DMVA mechanical systole appears to match that in the native beating heart states, the late systolic downward slope during DMVA appears to decline much more rapidly in all flow categories.
DISCUSSION
This study demonstrated the ability of DMVA support of the fibrillating heart to generate pulsatile flow similar to the native beating heart in mature canine and porcine animal models. The near-normal flow condition (75%+ of the baseline) represents the optimal performance of both DMVA support and ROSC relative to baseline. Further comparisons were examined in the moderate and severe shock
119
categories to demonstrate the effect of mean hemodynamics on diverse pulsatility parameters. Previously reported pulsatility parameters are based on the assumption every waveform can be broken into continuous and pulsatile components. The ratio between these two components is used to assess relative pulsatility.
Pulse pressure remains the most intuitive method for quantifying pulsatility since it can be readily visualized graphically. However, this method makes use of only two data points, thereby making it more vulnerable to underdamping effects that accentuate peak values (Romagnoli et al., 2011), and does not take the energy of the lifeforms into account by utilizing the whole waveform (Undar et al., 1999; Undar et al., 2005). Two strategies for measuring aortic pulsatility have become common, energy equivalent pressure and surplus hemodynamic energy (Souncy et al., 2013).
Energy equivalent pressure is calculated by dividing the integral of aortic power by the integral of aortic flow (effectively the work needed to pump a certain volume of blood) (Shepard et al., 1966). While EEP represents the total energy of the signal,
SHE represents the pulsatile component. Surplus hemodynamic energy is computed by subtracting MAP (continuous component) from EEP and converting to energy units with a conversion factor of 1,332. Pulsatile flow contains additional hemodynamic energy delivered to the vasculature that is not generated with diminished/none pulsatility. SHE provides insight in terms of energy that is utilized for pulsatile flow maintaining more physiologic microcirculation and better vital end-organ recovery than continuous flow.
120
Frequency domain analysis has also been used to assess pulsatility. A fast Fourier transform was used to generate the frequency spectrum of aortic flow, which allows for analysis of each harmonic in the frequency domain. Pulsatility index compares the amplitude at 0 Hz (representing the continuous component) to the sum of the harmonic amplitudes (representing the pulsatile component) (Kawahito et al., 2000).
The first five harmonics were used for PI calculations in this study. In this case PI is not to be confused with the Doppler-derived calculation of the same name that compares systolic and diastolic blood velocities in peripheral blood vessels
(Hashimoto et al., 2010). PI indicates the relative sharpness of a given waveform with respect to its mean flow, which is a valuable index for assessing the pulsatility.
Pulse power index is a slight modification of PI where harmonic amplitudes are multiplied by their respective angular frequencies (ω) before summation (Kaebnick et al., 2007). Using PPI allows for quantification of the relative power of a pulsatile waveform with respect to a non-pulsatile equivalent flow. Notably, only flow waveforms were analyzed in this study as the harmonic amplitudes have more significant amplitudes relative to the 0 Hz value than pressure.
Hypoxic ischemic brain injury has been known as the leading cause of mortality and dismal outcomes after cardiac arrest (Gräsner et al., 2016). Neurologic outcomes are determined by primary injury from immediate cessation of cerebral blood flow during CA. The principal goal of cardiopulmonary resuscitation is to
121
early return sufficient cerebral blood flow and oxygen delivery, and therefore optimize the neurological preservation after cardiac arrest and ROSC. During optimal DMVA support (defined as 75%+ baseline), mean flows (80.7% baseline) were significantly lower compared to both baseline and ROSC, while mean pressures did not significantly differ from the other two states. DMVA support effectively provides 93.4% and 78.6% cardiac output compared to the native beating heart of canine and swine respectively (Table suppl. 1). This phenomenon has been observed in direct cardiac compression devices and does not impede devices in properly supporting end-organ function (Oz et al., 2002; Artrip et al., 2000). In addition, relatively low flow during resuscitation may contribute to protection against the secondary injury to the brain including reperfusion injury, hyperoxia, hyperthermia, and microcirculatory dysfunction (Sekhon et al., 2017).
Besides beneficial effects on returning adequate cardiac output, DMVA support was able to achieve near-normal (close to 100% baseline) pulsatility (PP, EEP, SHE, PI,
PPI) compared to the physiological state of the native beating heart (Figure 31).
DMVA also parallelly exhibited superior pulsatility (PP, SHE, PI, and PPI) when compared to ROSC at all three flow conditions. Given the previously discussed benefits of pulsatile flow on neurological recovery (Anstadt et al., 1990; Anstadt et al., 1991c; Anstadt et al., 1992; Anstadt et al., 1993), a case can be made that animals in these studies would have benefitted from continued DMVA support during these unsupported post-resuscitation periods. The results of this study
122
demonstrate several pulsatility parameters (SHE, PI, and PPI) increase during
DMVA as hemodynamics (pressure and flow) deteriorate (Table 8 and Figure 31B).
Therefore, these previously-described pulsatility indices do not reflect declines in hemodynamics as they appear to be accentuated in states of shock and may be affected by compensatory physiology (Li et al., 2010). Furthermore, this inverse relationship between hemodynamics and pulsatility implies that DMVA, as an external energy provider, produces more enhanced vascular pulsatility when native circulatory function exacerbates with depletion of preserved internal aortic power.
In this study, aortic power waveforms were analyzed by incorporating all the important elements of pulsatility (pressure, flow, and time), thus directly assess the nature of the pulsatile waveform. These curves better demonstrated similarities in pulsatility between the baseline beating heart and DMVA support of the fibrillating heart. During optimal DMVA support, aortic power curve at early mechanical systole (upslope of the curve) appears to be similar compared to the same phase of the native cardiac cycle. DMVA generating aortic power, however, appeared to decline faster during end-systole. This phenomenon, observed on both canine and swine models (Figure suppl. 3) , may be caused by the difference of systole-to-diastole (S/D) switching between DMVA and native beating heart.
Intra-myocardial calcium transient decay contributes to the weakening contractile strength and induction of myocardial relaxation. The duration of late systole is determined by the rate of calcium transient decay. This natural process may occur
123
slower than mechanical S/D switching generated by the piston displacement of the
DMVA drive system. The finding is considered as a characteristic behavior of
DMVA support in preparation for active diastolic assist, and may account for lower mean pressures and flows. The average aortic power waveforms also demonstrated a marked attenuation in peak power with ROSC compared to baseline or DMVA support in both the near-normal and moderately-depressed flow states.
Amplification of the peak power region during DMVA is likely the driving factor behind the enhanced pulsatility seen during DMVA support.
VF was considered an important condition for testing given its common occurrence in the resuscitation setting and absence of meaningful contractility. This provides a clinically relevant scenario of CA where hemodynamics resulting from DMVA support are generated exclusively by external mechanical forces. The near-physiological pulsatility observed during DMVA support coupled with previous data suggesting superior neurologic recovery with DMVA compared to
CPB (Anstadt et al., 1990; 1991c; 1992; 1993) shows the potential of external actuation during resuscitation. While different settings were tested throughout this study due to variation in baseline cardiac physiology, external cardiac assist with proper device tuning is capable of replicating physiologic pulsatility for superior resuscitative support and perfusion of the brain in the absence of adequate native cardiac function.
124
This study illustrates promising resuscitative results for DMVA but does have some limitations. First, the current study is using an animal model of VF induced cardiac arrest without preexisting cardiac diseases, which may not comprehensively simulate CA patients in the clinical situation. Additionally, the effects of DMVA on hemodynamics were recorded at different flow conditions to optimize the setting of our resuscitative device. Based on the aortic power waveforms, mean hemodynamics during DMVA could potentially be improved by increasing maximum air pressure or slowing down retraction during end-systole. While these corrections would make the power waveform better match baseline, it is unknown how this might affect the active diastolic assist. Finally, Administration of neuroprotective agents and other complimentary treatments during DMVA support may further improve neurologic recovery, which is warranted to investigate in the future research.
CONCLUSION
This study demonstrates DMVA support of the fibrillating heart results in pulsatile flow characteristics similar to the native beating heart. Pulsatility during DMVA support of the fibrillating heart exceed those generated by the native beating heart following ROSC. The enhanced pulsatile energy generated by DMVA may provide ideal characteristics for resuscitative cerebral perfusion and neurological recovery.
125
Additionally, aortic power waveforms provide a more objective means for analyzing pulsatility. This methodology should be adopted in future analyses of pulsatile flow characteristics to better determine what aspects of curves are of physiologic importance.
126
Figure 30. Experimental design. Three states were studied: baseline, DMVA support during fibrillation, and recovery after ROSC. Repeated fibrillation cycles generated increasingly depressed recovery hemodynamics. Comparisons were made between groups according to mean flows during DMVA and ROSC relative to baseline (75%+: normal flow, 50-75%: moderate shock, and 25-50%: severe shock).
127
Table 7. Estimates of pulsatility used in this study.
Measure Equation Description
Pulse pressure is the most basic estimate, Pulse Pressure PP = Psys − Pdias but suffers excessively from artifact (PP) since only arterial pressure range is used.
Energy equivalent pressure (EEP) is the
Energy integral of aortic power (work) divided
∫(퐹푙표푤×푃푟푒푠푠푢푟푒)푑푡 Equivalent EEP = by the integral of aortic flow (volume ∫ 퐹푙표푤 푑푡 Pressure (EEP) pumped). This measure represents the
total energy of the waveform.
To estimate the pulsatile energy
Surplus component, mean arterial pressure
Hemodynamic SHE = 1,332(EEP − MAP) (MAP) must be subtracted from EEP. A
Energy (SHE) conversion factor of 1,332 is applied to
convert to energy units.
Fast Fourier transforms of aortic flows Pulsatility Index ∑푛 퐴 2 푖=0 푖 were calculated. The ratio of the DC (0 PI = 2 (PI) 퐴0 Hz) amplitude (A ), representing a flat 0
average, to the combined amplitudes of 2 ∑푛 휔 2퐴 Pulse Power 푖=0 푖 푖 the harmonic frequencies (the first ten, PPI = 2 퐴0 A ) representing a series of sinusoids Index (PPI) 1-10 (휔: 푎푛푔푢푙푎푟 푓푟푒푞푢푒푛푐푦) fitted to the periodic flow waveform.
128
Table 8. Hemodynamic and pulsatility measurements during baseline, DMVA during VF and ROSC post support. Three flow conditions represent DMVA resulted in 75%+, 50-75%, and 25-50% of the baseline aortic flow. Values expressed as mean ± SEM; a p<0.05 vs. Baseline, b p<0.05 DMVA vs. ROSC in same flow group. CO, cardiac output; MAP, mean arterial pressure; PP, pulse pressure; EEP, energy equivalent pressure; SHE, surplus hemodynamic energy; PI, pulsatility index;
PPI, pulse power index.
129
130
Figure 31. Comparisons of all the normalized hemodynamics and pulsatility measures included in this study. (A) CO, MAP, PP, EEP, PI, and PPI in the near-normal flow condition. (B) SHE in all three flow categories. DMVA support provided pulsatility (evaluated by PP, EEP, SHE, PI, and PPI) similar to 100% baseline when aortic flow was near normal. SHE increased during DMVA as
131
hemodynamics deteriorate from normal flow to severe shock. CO, cardiac output;
MAP, mean arterial pressure; PP, pulse pressure; EEP, energy equivalent pressure; PI, pulsatility index; PPI, pulse power index; SHE, surplus hemodynamic energy.
Values are expressed as mean ± SEM; * p < 0.05, vs. Baseline (same flow); † p <
0.05, vs. DMVA (same flow); • p < 0.05, vs. DMVA in normal flow; • • p < 0.05, vs.
DMVA in moderate shock.
132
133
Figure 32. Aortic power waveforms for three experimental states at different flow levels. Aortic power waveforms were calculated over a standard 700 millisecond cardiac cycle for captures in all three flow groups. Instant aortic power is equal to the product of simultaneous aortic flow and pressure. An ensemble average was used to generate a characteristic curve for baseline, DMVA support, and
ROSC in all three flow groups. Standard error bars are presented to demonstrate signal variation.
134
Table suppl 1. Pulsatile flow characteristics of canine and swine at normal flow level (75%+ baseline flow). Values expressed as mean ± SEM; a p<0.05 vs.
Baseline, b p<0.05 vs. DMVA. ROSC rate includes all animals achieving ROSC, no matter what level of flow. ROSCnf rate only includes animal achieving ROSC at normal flow (nf) level.
Canine (n=11) Swine (n=5)
Baseline DMVA ROSC Baseline DMVA ROSC Subject # 11/11 9/11 6/11 5/5 3/5 3/5 73.3% ROSC rate - - - - 55.6% (5/9) (11/15) ROSCnf rate - - 40.0% (6/15) - - 33.3% (3/9)
a a CO (L/min) 3.32 ± 0.10 3.10 ± 0.06 3.20 ± 0.07 3.08 ± 0.14 2.42 ± 0.08 2.32 ± 0.12
a b MAP (mm Hg) 74.2 ± 4.0 70.5 ± 2.2 68.8 ± 2.7 71.7 ± 1.9 48.8 ± 0.7 68.0 ± 1.9
b PP (mm Hg) 67.8 ± 6.6 71.4 ± 3.6 50.6 ± 4.4 37.2 ± 4.3 36.9 ± 1.5 27.6 ± 4.3
a b EEP (mm Hg) 92.1 ± 3.8 88.4 ± 2.1 82.0 ± 2.6 91.7 ± 2.7 56.1 ± 1.5 84.3 ± 2.4
23,945.6 24,813.9 21,432.5 20,651.9 15,156.5 16,817.4 3 SHE (ergs/cm ) ±2,337.7 ±1,330.5 ±1795.2 ±2,768.2 ±1,531.6 ±2397.3
a a PI 0.27 ± 0.04 0.37 ± 0.02 0.29 ± 0.03 0.94 ± 0.09 0.43 ± 0.05 0.24 ± 0.08
a ab PPI 1.34 ± 0.25 1.99 ± 0.14 1.43 ± 0.20 2.36 ± 0.19 1.46 ± 0.11 0.71 ± 0.17
135
136
Figure suppl 3. Aortic power waveforms of canine and swine. Both canine and swine aortic power waveforms were separately calculated over a standard 700 millisecond cardiac cycle for captures in all three flow groups. Instant aortic power is equal to the product of simultaneous aortic flow and pressure. An ensemble average was used to generate a characteristic curve for baseline, DMVA support, and
ROSC in all three flow groups. Standard error bars are presented to demonstrate signal variation. DMVA has noticeably higher peaks than ROSC in canine subjects.
DMVA support of canine subjects produced a much steeper downslope in late systole versus baseline and ROSC, while ROSC had a significantly blunted peak region compared to either baseline or DMVA support. Power during DMVA support declined substantially from baseline in both moderate and severe shock.
137
CHAPTER VII Echocardiographic Characteristics of a Mock Ventricle are Similar to the Fibrillating in vivo Ventricle during Direct Mechanical Ventricular Actuation
This chapter addresses SA 3 of this dissertation.
Specific Aim 3: Characterize mock LV mechanics during DMVA support in contrast to in vivo LV mechanics of the supported fibrillating heart.
138
RATIONALE
Direct mechanical ventricular actuation (DMVA) is a non-blood-contacting ventricular compression device that augments both systolic and diastolic pump function for mechanical support of the fibrillating or failing heart. (Anstadt et al.,
1995; Perez-Tamayo et al., 1995; Anstadt et al., 2009; McConnell et al., 2014; Zhou et al., 2018a; 2018b). Notably, DMVA was recognized by the American Heart
Association for its potential in resuscitative circulatory support of patients with cardiac arrest (Cummins et al., 2003).
Development of an acute mechanical circulatory support (MCS) device such as
DMVA requires many repeated trials on different experimental subjects including mock circulation loops (MCL) that are mechanical representations of the heart and circulatory system. In vitro assessment of a mechanical circulatory support device in an MCL is frequently utilized to refine device designs before expensive in vivo testing. Successful evolution of an MCL that simulates the physiology of the cardiovascular system ensures the evaluation of the device’s capability to maintain adequate hemodynamics under pathological conditions (Pantalos et al., 1998; Lee et al., 2009).
A complete mock circulation system consists of cardiac and vascular components assigned mechanical properties and enables reproduction of desired hemodynamic
139
characteristics (Papaioannou et al., 2003; Timms et al., 2005). In order to assess ventricular mechanics of the DMVA supported mock heart which is comparable to myocardial mechanics measured from animal subjects during DMVA support, speckle tracking echocardiography (STE) was used in both in vitro and in vivo systems. Our laboratory has been developing a biventricular mock circulatory system (BMCS) which meets all criteria stated above, to evaluate the DMVA device in vitro. The purpose of this study was to determine the comparability of the BMCS to a large animal cardiac arrest model by comparing ventricular mechanics during mechanical support of mock ventricles and akinetic porcine hearts.
RESEARCH DESIGN
Resuscitation Swine Model
The Wright State University Lab Animal Care and Use Committee approved the experimental protocol, and all animals were treated in compliance with the Guide for the Care and Use of Laboratory Animals of the National Academies, published by the
National Institutes of Health, revised 2011.
Adult Yorkshire swine (n=8, 80±2kg) were selected for heart sizes similar to the average adult human hearts and studied using a model of repeated periods of cardiac arrest (Figure 33). Animals were anesthetized (1-2% isoflurane), underwent median
140
sternotomy, pericardiotomy, and instrumented for hemodynamic monitoring (Chapter
III, surgical protocol). Following baseline recordings of the native beating heart, ventricular fibrillation (VF) was induced using a 9-volt electronic shock for 5 minutes of unsupported circulatory arrest. DMVA was then applied for 15 minutes followed by defibrillation attempts. After return of spontaneous circulation (ROSC), hemodynamics was monitored for 15 minutes. Other than the administration of standard doses of lidocaine, no other cardiovascular agents such as vasopressors or inotropes were administered during the experimental protocol. Intravenous fluids were administered to maintain central venous pressures in the physiologic range.
Animals were euthanized by potassium chloride injection (1-2 mmol/kg) under anesthesia at the end of the experiment (Beaver et al. 2001).
Mock Circulatory System Set-up
A biventricular mock circulatory system was utilized to evaluate ventricular mechanics during assist (Zhou et al., 2017a; 2017b). As a complete circulatory loop, this BMCS system comprised of a heart substitute (mock bi-ventricles and atrial reservoirs connected by atrioventricular (AV) platform) and vasculature components
(systemic and pulmonary resistance, and compliance elements) (Figure 34).
Vasculatures of systemic and pulmonary circulation were based on the four-element
Windkessel model, consisting of aortic valve resistance (characteristic impedance), peripheral resistance, total arterial compliance, and aortic inertance (Timms et al.,
2011; Segers et al., 2008). Compliance was regulated by altering the fluid level in the
141
pertinent reservoirs. Resistance was adjusted with gate valves. The inertance element was accounted for using these adjustable components and taking in to account for fluid viscosity.
Silicone mock ventricles with similar material and anatomical properties to the native swine heart were tested with the same DMVA device (Figure 35). Using anatomical measurements from extracted swine hearts, a computational design was developed in the SolidworksTM software to 3D-print a ventricular mold (including intact left and right ventricles, and septal wall). The mock ventricles were integrated with the AV platform which is attached to two atrial reservoirs. The pulmonary valve and tricuspid valve were seated in the base of the AV platform providing minimal dead space between the valves and the right ventricle (RV). The aortic and mitral valves were positioned above the AV platform (Figure 34). Silicone mock ventricles were firmly clamped over the inner mounting flange of the AV platform.
DMVA Support Setting
DMVA support rates were adjusted to match natural rhythms in animal subjects.
Subjects had heart rates approximately 80 bpm. Systolic actuation air pressures were adjusted to produce baseline systolic blood pressures and diastolic air pressures were adjusted to fully retract the inner membrane to the semi-rigid outer shell to facilitate diastolic assist. Swine, as well as the mock ventricles used 105-100 mm diameter cups
(measured at the rim of the cup near the AV groove). A low intensity apical vacuum
142
suctioned the heart in the cup and generated a tight interface between the epicardial surface and device inner membrane.
Speckle Tracking Echocardiography
Intracardiac echocardiography was performed on the swine model using an
ACUSON Sequoia System (Siemens Medical Solutions, Mountain View, CA) with a 10-MHz Acuson AcuNav10F intravascular catheter (Siemens Medical Solutions,
Mountain View, CA) positioned in the right atrium. B-mode cine clips (60 Hz) were obtained and image processing were performed offline using Syngo Vector Velocity
Imaging software (Siemens Medical Solutions, Mountain View, CA).
B-mode clips with adequate visualization of the endocardial border were selected.
The endocardial LV borders were manually traced and accurate tracking verified >3 cycles in the four-chamber view at end-diastole to compute Lagrangian strain and derivative strain rate of the LV. At least five endocardial border tracings were performed in each ECHO clip, and a total of 80 swine clips were compared to 80 mock clips. End-systolic and end-diastolic longitudinal strain values for every segment of the LV (basal/mid/apical septal wall, and basal/mid/apical lateral wall) were recorded and averaged to obtain global longitudinal strains (GLS) for each trial. Peak strain rates of global and regional LV wall, derived from the above strain data, were also analyzed in this study.
143
Mock ventricles were instrumented, and the same echocardiographic probe was positioned right above the chamber through the AV platform in order to obtain comparable echocardiographic images and video clips. Same imaging system and software were utilized for the mock LV mechanics analysis. Boarder of the inner mock LV wall were accurately traced in the longitudinal two-chamber view at end-diastole to calculate Lagrangian strain and strain rate at the base, mid, and apex of the LV. Mock LV mechanics profile also consists of systolic and diastolic longitudinal strain and peak strain rates for global and six regional walls.
Data Collection and Statistical Analysis
A TS420 perivascular flowmeter (Transonic Systems Inc, Ithaca, NY) was connected to an ultrasound probe placed on the ascending aorta to measure porcine cardiac output (CO). A P23XL transducer (Spectramed Inc, Mt Vernon, OH) was introduced into the aorta via the right carotid. Aortic pressures and flows were periodically collected throughout the experiments for 10 sec intervals at 200 Hz using LabVIEW data acquisition software (National Instruments, Austin, TX). Intracardiac echocardiography was performed to obtain ventricular geometry and mechanics profile. Each animal required at least one successful defibrillation attempt to be included in the final analysis.
144
Simulated hemodynamics and intracardiac echocardiography in the mock circulatory system were collected using the same flow, pressure, and intravascular ultrasound transducers describe above.
Statistical analysis was performed using JMP version 13 commercial software (SAS
Institute, Cary, NC). All values are expressed as mean ± standard error of the mean
(SEM). Normality tests were first carried out. For normally distributed data, a two-sided, unpaired t test was used for comparing between two models, with an α level of 0.05. Regional LV wall strains and peak strain rates were compared using
1-way ANOVA with post hoc Tukey’s HSD tests.
RESULTS
In this study, we compared echocardiographic and hemodynamic characteristics between an in vivo porcine model (n=8) and an in vitro mock model (n=8). These two models utilized similar DMVA support to superimpose mechanic features on the akinetic LV. The mock ventricle consists of both left and right ventricle, which provided proper interface for DMVA cup and mock circulatory system. The biventricular mock resembled the natural LV anatomical silhouette and dimensions observed via intracardiac echocardiography (Figure 36). It is noteworthy that
B-mode images showed the structure of the DMVA supported mock bi-ventricle similar to the supported swine heart at end-diastole and M-mode captures displayed
145
clear and complete cardiac dimensions of the mock bi-ventricle similar to the native heart (Figure 36C).
Native and mock LV Volume Analysis
We compared LV volumes of the native beating heart and DMVA supported VF arrested heart (Table 9). End-diastolic volume (EDV) and end-systolic volume
(ESV) both showed no significantly difference between two models. Consistently, stroke volume (SV) of the mock LV was similar to the fibrillating heart during
DMVA support (P=0.053), considering that SV is the difference of EDV and ESV.
In addition, pump function indicator ejection fraction (EF) during DMVA support was compared and observed no significant difference between models.
Hemodynamics Evaluation
Bi-ventricular mock circulatory system provided a hypodynamic condition similar to the resuscitated porcine heart during DMVA assist, although systolic blood pressure of the mock model was significantly lower than the in vivo model (Table
9). Aortic flow was detected similar during DMVA support of the fibrillating heart and mock ventricles.
146
Speckle Tracking Echocardiographic Assessment
DMVA assist of the mock bi-ventricle produced GLS during either systole or diastole similar to the supported VF porcine heart (Table 9), which was consistent with findings of LV volume comparison (ESV and EDV respectively). Spatial distribution of regional longitudinal strain (RLS) over six segments of the LV
(native and mock) at end-systolic and end-diastole was shown in heat maps (Figure
37). LV all six segments (basal septal, mid septal, apical septal, apical free, mid free, and basal free walls) in the apical long-axis plane are colored from green to red to represent absolute values of RLS from higher to lower ones. In general, mock LV during DMVA support manifested polarized RLSs for six segments (color variance was wider) at either end-systole or end-diastole. Swine model, on the other hand, exhibited more balanced RLSs for six segments (color variance was closer) at end-systole and end-diastole. Six segments of LV wall were comprehensively studied. We found that DMVA support of the mock LV produced both end-systolic and end-diastolic longitudinal strain of each of the six regions similar to those observed in the DMVA assisted fibrillating heart (Figure 38 and Table suppl. 2).
Intra-group comparison revealed that RLS of the apical septal wall was significantly higher than basal free wall during systole in the swine model (Figure suppl. 4A).
RLS of the apical septum of the mock LV was significantly enhanced during systole and diastole when compared to basal free wall in the mock model. Furthermore, apical lateral RLS was significantly reduced versus apical septum during systole only found in the DMVA supported mock LV (Figure suppl. 4B).
147
Notably, LV peak global longitudinal strain rate (GLSR) over both systole and diastole was significantly reduced during DMVA support of the mock bi-ventricle when compared to the supported fibrillating heart (both Ps < 0.001) (Table 9 and
Figure 39). LV lateral peak regional longitudinal strain rates (RLSR) were significantly lower in mock ventricles versus supported fibrillating heart (all three Ps
< 0.01), yet regional septal peak strain rates were similar between mock and swine (all three Ps > 0.05). The above phenomenon was observed in both systole and diastole
(Table suppl. 3). Intra-group comparison demonstrated that apical septum of either native or mock LV during DMVA support showed elevated RLSR comparing to other five LV segments throughout the cardiac cycle (Figure suppl. 3). However, porcine apical septal RLSR was significantly higher than other two septal segments during
DMVA assist; mock apical septal RLSR, on the other hand, was significantly superior than all three lateral RLSRs.
DISCUSSION
The aim of this study was to demonstrate the capability of the BMCS to reproduce clinically relevant ventricular mechanics characteristics during mechanical circulatory support. This study’s results elaborated, first of all, the BMCS’s capacity to entirely reproduce the akinetic hearts being supported by DMVA, in the presence of hypotension which was parallel with the hemodynamic condition of the mechanically
148
supported porcine model. Insignificantly different aortic flow was generated when
DMVA was applied to the BMCS platform versus in vivo VF resuscitation model.
This complete mock loop simulates arterial hemodynamics using a four-element
Windkessel model (WK4), consisting of peripheral resistance, total arterial compliance, aortic valve resistance (that is, characteristic impedance), and aortic inertance (Timms et al., 2011; Segers et al., 2008).
WK4 is a lumped-parameter model, which lumps the distributed properties of the arterial circulation into discrete number of parameters when neglecting continuous wave propagation and reflection (Westerhof et al., 2009). It is not suitable for evaluating spatially distributed phenomena and aspects of wave travel, but it simply yet accurately represents the approximation of ventricular afterload. The resistive element reflects the resistance generated by small arteries and arterioles, while the compliance element mimics the elastic and buffering features of the large arteries such as aorta. Herein, via modifying resistance and compliance components, a reproducible physiological or pathophysiological model is ready for dynamic characterization of external ventricular actuation.
The value of this mock platform lies in its 3D-printed biventricular mock heart, hands-on robustness, and its four-element Windkessel design, allowing for the LV geometrical adjustment and interchange of MCS devices. As such, it provides an interactive testbed, affording the ability to evaluate hemodynamic effects and
149
consequences of pressure and volume variations encountered in patients with cardiac arrest and acute MCS devices. The fibrillating heart during cardiac arrest is essentially in a non-contractile state and provides akinetic LV wall motion without external mechanical support. Silicone mock ventricles due to their hardness
(resistance to permanent indentation) and stretch properties, was proved to serve as an eligible surrogate of the fibrillating heart (Gregory et al., 2009).
Comparing to commonly used pneumatic artificial ventricles, which are controlled by the external drive console (Sénage et al., 2014; Crosby et al., 2015; Crosby et al., 2017), 3D-printed silicone ventricles has the intrinsic advantage of mimicking the shore hardness of native myocardial tissues and flexibly modifying anatomical structures (including stiffness, thickness, and integrity of the myocardial wall) for different cardiomyopathies. In this study, we utilized silicone with its shore hardness of OO-20 (Young’s modulus value was approximately 0.45 MPa) to fabricate the mock bi-ventricle, as this type of chamber replicates the shape and flexibility of the natural ventricle. Mock LV geometry analysis revealed no significant difference when compared to the porcine LV in terms of EDV, ESV, and SV (Table 9). Such an elastic chamber provides the ability to test ventricular mechanics during DMVA support, which would be valuable and comparable to animal or human data in the similar setting of DMVA assist. Notably, no pneumatic artificial muscle was used given that the mock heart was designed to simulate a fibrillating, non-contractile heart. For evaluation of long-term ventricular assist devices (VAD) applied to
150
patients with chronic heart failure, a pneumatic muscle (mimicking heart’s striated muscle fibers) actuated silicone LV can be optimal to establish an in vitro heart failure model, especially when these robotic muscles enable the spiral contraction of the mock heart (Baturalp et al., 2015; Roche et al., 2017). Furthermore, apical cannulation of the mock ventricle can be easily achieved when studying blood-contacting VADs.
This study utilized echocardiography to evaluate silicone mock LV wall mechanics during DMVA support. The results indicated that 2-D speckle-tracking echocardiography can be used to characterize the strain properties of mock silicone ventricles via an intracardiac interrogation (ultrasound probe positioned right above the mock LV). The mock ventricle was thereby visualized in the basal two-chamber echocardiogram view, which was comparable to the animal heart imaged in a similar basal four-chamber view (Figure 36). The strain rate and strain magnitudes of the mock LV demonstrated good correlation during mechanical systole and diastole with those obtained from a fibrillating heart during in vivo experiments. Specifically, GLS and RLS (all six segments) of the DMVA supported mock LV were similar to the
DMVA supported fibrillating swine LV. These findings were true for both systolic and diastolic strain comparisons (Figure 38). These between-group analyses indicated similarities in systolic strain behavior when comparing all six segments in the swine
LV versus the mock LV during DMVA support. Intragroup analyses of regional systolic LV wall motion relationships revealed further evidence of such similarities.
151
Specifically, the systolic apical septal regional strain was found to be significantly higher than the systolic basal free wall strain in both the fibrillating swine LV and the mock LV during DMVA support. While diastolic strain in all six regions were similar between the swine fibrillating LV and the mock LV, intra-group analyses revealed some differences. Specifically, diastolic regional strain in the apical septum was significantly elevated when compared to basal free wall of the mock LV but were similar in the swine LV (Figure suppl. 4). Therefore, although the systolic characteristics of the mock LV appeared to mimic that of the fibrillating swine LV, diastolic characteristics appeared to be different with respect to these intra-group analyses.
Peak strain rate describes the maximal rate of tissue or material deformation. This has been used as a non-load-dependent measure of myocardial function in the beating heart. Therefore, peak strain rates were used as a non-load-dependent measure for comparing LV mock dynamics to that of the fibrillating swine heart in the study. Results indicate that global systolic strain rates were significantly decreased in the mock model when compared to the fibrillating swine heart during
DMVA assist. However, when evaluating regional strain rates, it was found that septal regional peak strain rates were similar between the two groups (Figure 39).
The discrepancy between the LV free wall and septum counted for the reduction in global systolic strain rates when comparing the mock LV to the fibrillating swine heart. Wall thickness or hardness is the potential explanation for these findings. It is
152
noteworthy that when using intra-group analyses, the apical septum appeared to manifest the greatest LV strain rate among all the LV segments in both the mock and swine model. This finding was true for both systole and diastole (Figure suppl.
5). Therefore, the effect of DMVA support on the mock LV appeared to have similar consequences with respect to LV regional wall relationships as it relates to peak strain rates.
Importantly, this unique BMCS may be utilized for inter-ventricular actuation device comparative studies and research, circumventing the prohibitive cost and ethical issues of invasive MCS device implantation on the human subject. While we report only on the BMCS as a model of normal-sized akinetic bi-ventricle, from ongoing studies by our group, it is explicit that the platform can physically emulate over-sized ventricles (i.e. dilated cardiomyopathy) as well as interventricular septal defect, allowing universal echocardiographic and hemodynamic characterization to be performed and inter-device comparative studies to be undertaken.
DMVA aims to help heart physiologically pumping (ejecting and relaxing at a fixed systolic/diastolic duration ratio of approximately 0.7) and restore the cardiac pump function and pulsatile hemodynamics post VF cardiac arrest. Besides it was investigated as an acute MCS device, DMVA offers effective circulatory support of the failing heart (Anstadt et al., 2009; McConnell et al., 2014). In order to evolve to an ideal MCS device, DMVA eventually should have the capacity and durability to
153
maintain normal cardiac output both at rest (5-6 L/ min) and during exertion (15-20 L/ min); supporting without blood contact; and being non-obligatory (the device is able to be turned on and off as needed). The long-term development stage prior to clinical use needs discipline and resilience, but most of all, funds. Generation of a qualified complete mock circulation loop to assess the device can promote the development of DMVA in a cost-efficient path (without a strict dependence on animal trails for testing). This novel BMCS incorporating a compliant 3D-printed anatomical bi-ventricle is an accurate mechanical representation of the heart and vascular system, which also has the potential to constantly undergo adjustments/improvements and reduce the number of animal trials required.
This study highlighted the unique features of this BMCS including the silicone mock bi-ventricle and characterized both mock and natural LV mechanics during DMVA support. However, it does have some limitations. First, using ultrasound to characterize strain changes in synthetic ventricles still need to be verified carefully.
There might be different ultrasound responses for silicone versus myocardial tissues, which are expected be overcome using modified speckle tracking algorithm once we get to know the conversion relationship between the two matters. Moreover, no radial or circumferential deformation data was collected due to the ultrasound interference of DMVA support in the B-mode clips. High frame rate ultrasound probe might be the solution in future studies. Short-axis view of the echocardiography offers informative circumferential deformation data to weigh in cardiac twisting and untwisting during
154
systole and diastole respectively. Finally, assessment of DMVA cup surface dynamics could potentially reveal relationships of the ventricular boundary profile and ultrasound results, which is warranted in future studies.
CONCLUSION
This study demonstrated the mock ventricles, sharing similar dimensions of both ventricular chambers, echocardiographically resembled the natural anatomical silhouette of swine ventricles. A hypotensive mock circulation with relatively normal blood flow was produced in consideration of being comparable to the lower-level hemodynamics post-sternotomy observed in the porcine model. DMVA support generated similar GLS and RLS throughout the cardiac cycle in both animal and mock LV. Furthermore, peak strain rate of the mock ventricle was significantly reduced during DMVA assist when compared to the supported fibrillating heart. In this work, STE was utilized as an effective means for evaluating the deformation of both myocardial tissue and silicone rubber. Quantitation of strains and strain rates in the mock ventricles provided an objective basis for comparison with the native heart in the context of DMVA support. Silicone mock ventricles have great potential to be surrogates to the native fibrillating hearts to evaluate ventricular dynamics and pump function during DMVA support, and this unique biventricular mock circulatory system is expected to provide a practical testbed for developing DMVA or any MCS devices utilizing the concept of external ventricular actuation.
155
Figure 33. Animal experimental design. DMVA resuscitation of the swine (yellow) was the experimental state compared with same-sized mock ventricles during DMVA support in this study. Baseline data was collected of the native beating heart. Animals then were subjected to 5-minute of ventricular fibrillation with total circulatory arrest followed by DMVA support. This design allows 3-5 arrest/recovery cycles to test
DMVA during VF.
156
157
Figure 34. Schematic of the complete mock circulatory system with biventricular mock heart attached. P: pressure transducer; Q: flow meter; R: resistance valve; LA: left atrium reservoir; RA: right atrium reservoir; LV: left ventricle; RV: right ventricle; AoV: aortic valve; MV: mitral valve; PV: pulmonary valve; TV: tricuspid valve. Black block: mechanical heart valve (4 in total). Solid line: systemic circulation; Dotted line: pulmonary circulation. This BMCS was built to efficiently evaluate the DMVA device for multiple circulatory simulations at different hemodynamic settings.
158
Figure 35. Comparable swine and mock ventricles. (A) Extracted swine heart;
(B) DMVA supported fibrillating heart; (C) Axial view of the mock ventricle fabricated using the anatomical dimension of a swine heart; (D) Longitudinal view of mock ventricles inside in the DMVA cup. LV: left ventricle; RV: right ventricle.
159
Figure 36. Mechanical diastolic actuation (arrows denote DMVA housing) on the (A) fibrillating swine heart and (B) mock bi-ventricle. (C) M-mode displaying cardiac dimensions of the mock bi-ventricle. Ultrasound probe was positioned at right atrium for the swine model and above LV for the mock model, which leads to opposite location of LV and RV shown in the echocardiographic image. Note the similarities between DMVA supported fibrillating heart and the mock ventricles. LV: left ventricle; IVS: interventricular septum; RV: right ventricle; S: systole; D: diastole.
160
Table 9. Echocardiographic and hemodynamic comparison between swine and mock model. Values expressed as mean ± SEM; Results compared using unpaired t-tests; EDV: end-diastolic volume; ESV: end-systolic volume; SV: stroke volume;
EF: ejection fraction. * p < 0.05, vs. fibrillating heart; * * p < 0.01, vs. fibrillating heart.
Fibrillating Mock Heart P Value Heart LV Volume(ml) EDV 64.3±1.6 58.7±2.3 0.054 ESV 34.6±1.5 32.3±2.2 0.400 SV 30.5±1.5 25.8±1.3 0.053 EF(%) 48.3±1.9 45.8±2.4 0.410 LV Global Longitudinal Strain(%) End-Systolic -14.0±0.9 -14.2±1.2 0.870 End-Diastolic 14.2±1.0 12.4±1.3 0.263 LV Global Peak Strain Rate(s-1)
Systolic -2.25±0.07 -1.52±0.09 ** <0.001 Diastolic 2.26±0.08 1.54±0.1** <0.001 Arterial pressure(mmHg) Systolic 73.4±1.4 56.4±3.6* 0.012 Diastolic 36.5±0.7 33.5±1.6 0.541 Aortic Flow(L/min) 2.75±0.07 2.95±0.18 0.292
161
162
Figure 37. Summary heat map of peak regional longitudinal strains (RLS, %) at both (A) end-systole and (B) end-diastole between swine and mock model.
Values are presented as mean. There were no significant differences for all six segments between two models. Visualized volume changes were generally consistent with the geometry profile demonstrated in Table 9 (ESV to EDV ratio is
53.8% for swine hearts and 55% for mock ventricles, respectively).
163
164
Figure 38. LV GLS and RLS comparison between animal and mock model.
SEM bars are shown. GLS: global longitudinal strain; RLS: regional longitudinal strain; SW: septal wall; LW: lateral wall. There was no significant difference between two models in terms of all strain measures.
165
Figure 39. LV mechanical inotropy and lusitropy reflected by peak strain rate.
Inotropy or lusitropy describes mechanical force and motion of the LV (native and mock) superimposed by DMVA support. Global and regional longitudinal strain rate
(RLSR) were compared between animal and mock model. SEM bars are shown.
GLSR: global longitudinal strain rate; SW: septal wall; LW: lateral wall.
166
Table suppl 2. Regional longitudinal strain (RLS) comparison between animal and mock model. Values expressed as mean ± SEM; Results compared using
Tukey’s HSD tests. Ns: No significant difference between two models (P>0.05).
End-Systole
Septal Wall Lateral Wall
Basal Mid Apical Basal Mid Apical
Swine -13±1.31 -11.72±1.32 -17.51±1.38 -11.15±1.36 -12.88±1.33 -14.38±1.33
Mock -15.24±1.67 -14.83±1.65 -17.81±1.74 -9.1±1.63 -13.82±1.62 -8.8±1.63
p-value ns ns ns ns ns ns
End-diastole
Septal Wall Lateral Wall
Basal Mid Apical Basal Mid Apical
Swine 13.03±1.42 11.99±1.41 11.1±1.47 14.06±1.47 14.54±1.41 14.32±1.41
Mock 8.32±1.79 9.16±1.84 15.54±1.82 7.6±2.16 10.5±1.77 9.31±1.82
p-value ns ns ns ns ns ns
167
Table suppl 3. Regional longitudinal peak strain rate (RLSR) summary of animal and mock model. Values expressed as mean ± SEM; Results compared using Tukey’s HSD tests. Ns: No significant difference between two models
(P>0.05).
Systole
Septal Wall Lateral Wall
Basal Mid Apical Basal Mid Apical
Swine -1.97±0.12 -1.81±0.12 -2.97±0.13 -2.10±0.12 -2.31±0.13 -2.41±0.13
Mock -1.50±0.16 -1.68±0.15 -2.38±0.17 -0.97±0.15 -1.46±0.15 -1.20±0.16
p-value ns ns ns <0.01 <0.01 <0.01
Diastole
Septal Wall Lateral Wall
Basal Mid Apical Basal Mid Apical
Swine 1.97±0.11 1.85±0.12 2.72±0.12 2.12±0.11 2.19±0.12 2.28±0.12
Mock 1.51 ±0.15 1.48±0.15 2.23±0.16 0.87±0.14 1.43±0.14 1.25±0.15
p-value ns ns ns <0.01 <0.01 <0.01
168
Figure suppl 4. Regional wall strain intra-group comparisons. (A) LV longitudinal strain during systole and diastole in the swine model; (B) LV longitudinal strain during systole and diastole in the mock model. Asteroid sign indicates that two compared regions were significantly different. GLS: global longitudinal strain; SW: septal wall; FW: free wall.
169
Figure suppl 5. Regional peak strain rate intra-group comparisons. (A) LV longitudinal strain rate during systole and diastole in the swine model; (B) LV longitudinal strain rate during systole and diastole in the mock model. Asteroid sign indicates that there was a significant difference between two compared segments.
GLSR: global longitudinal strain rate; SW: septal wall; FW: free wall.
170
CHAPTER VIII CONCLUSIONS AND FUTURE DIRECTIONS
This dissertation project tested the hypothesis that delivering mechanical forces during resuscitative support of the fibrillating heart can generate left ventricular dynamics and pulsatile hemodynamics similar to the physiological state. Direct mechanical ventricular actuation was used in this project as it has been shown to be an effective means for mechanical support of the arrested heart. DMVA utilizes a support cup that encompasses the ventricles and a drive system that provides pneumatic forces to the cup for its actuation and action on the ventricles. In this dissertation, optimal DMVA support was shown to mechanically affect the akinetic
LV in a fashion that resulted in LV pump function similar to the native beating heart.
Furthermore, DMVA support generated LV pump function dynamics and pulsatile hemodynamics which were similar to the physiological state. The near-physiological blood flow generated during DMVA support has previously been shown to benefit vital organ perfusion and neurologic outcome following cardiac arrest in laboratory studies (Anstadt et al., 1990; 1991c; 1992; 1993). Notably, DMVA uniquely provided diastolic support to the fibrillating LV which resulted in diastolic function and LV
171
myocardial deformation that mimicked that of the LV myocardium of the native beating heart. These findings may have important implications on recovery of post-arrest myocardial dysfunction following DMVA resuscitative support. The results pertaining to the mock circulatory system confirmed DMVA has similar effects on the akinetic mock silicone LV compared to the fibrillating swine heart.
Future directions
This dissertation provides valuable insights into functional manifestations of the VF arrested heart during DMVA resuscitative support. The device has the potential to result in LV pump performance during resuscitative support that mimics the native beating heart. Prior investigations have shown DMVA support for resuscitation has no significant deleterious effects on myocardial cellular integrity or pump function following ROSC. It would be valuable to explore mechanotransduction mechanism that may occur during DMVA cyclic compression and dilatation of the fibrillating heart. It is possible that these mechanical actions on the fibrillating heart may have important implications to myocardial functional recovery during ROSC. Areas pertinent to these questions would include the following. A) Will cyclic compression and stretch applied on an in vitro cardiac myocyte alter the function of important effectors involved in mechanotransduction (such as RyR2, SERCA2a, and titin
I-band)? B) What could possibly be the initial proteins or other molecular structures in sensing external mechanical forces within myocytes (maybe Nox2 on the T-tubule, or N2 domain on the sarcomere)? C) How about mechanosensitive components
172
existing in extracellular matrix (β1-integrins at costameres and focal adhesion plaques that connect to the cytoskeleton) and intercellular structures (calcium-dependent
N-cadherins at intercalated disks)? D) What are cellular responses in an in vivo heart model using mechanical support? A possible mechanotransduction theory could be made that mechanical forces (location, direction, and strength) are accurately sensed and delivered through a scaffold-like complex (from extracellular to intercellular to intracellular level) within myocardial tissues based on many previous findings scattered in relevant studies (Prosser et al., 2011; Prosser et al., 2011; Krüger et al.,
2009; Fukuda et al., 2010; Kresh et al., 2011; McCain et al., 2013). Notably, cytoskeletal integrity is mechanistically essential for mechanical stimuli transforming to cellular responses.
Additionally, the BMCS mock system can be useful for testing effects of DMVA on the arrested, fibrillating heart. Based on the comparability observed in silicone mock hearts versus the native heart, mock ventricles could be used to test at what conditions/configurations DMVA can generate optimal support to the cardiac dynamics and hemodynamics. In addition, mock ventricles could potentially be modified to simulate different cardiac pathological conditions.
173
CHAPTER IX REFERENCES
Aggarwal A, Pant R, Kumar S, Sharma P, Gallagher C, Tatooles AJ, Pappas
PS, Bhat G. Incidence and management of gastrointestinal bleeding with continuous flow assist devices. Ann Thorac Surg. 2012;93(5):1534-40.
Alzaga-Fernandez AG, Varon J. Open-chest cardiopulmonary resuscitation: past, present and future. Resuscitation. 2005;64:149-56.
Andersen OS, Smiseth OA, Dokainish H, Abudiab MM, Schutt RC, Kumar A,
Sato K, Harb S, Gude E, Remme EW, et al. Estimating Left Ventricular Filling
Pressure by Echocardiography. J Am Coll Cardiol. 2017;69(15):1937-48.
Anstadt GL, Blakemore WS, Baue AE. A new instrument for prolonged mechanical massage. Circulation. 1965;31(suppl II):43.
Anstadt MP, Anstadt GL, Lowe JE. Direct mechanical ventricular actuation: A review. Resuscitation. 1991;21:7–23.
Anstadt MP, Bartlett RL, Malone JP, Brown GR, Martin S, Nolan DJ,
Oberheu KH, Anstadt GL. Direct Mechanical Ventricular Actuation for Cardiac
Arrest in Humans. Chest.1991;100(1):86-92.
Anstadt MP, Budharaju S, Darner RJ, Schmitt BA, Prochaska LJ, Pothoulakis
AJ, Portner PM. Ventricular actuation improves systolic and diastolic myocardial function in the small failing heart. Ann Thorac Surg. 2009;88(6):1982-8.
174
Anstadt MP, Budharaju S, Wozniak CJ, Darner RJ, Prochaska LJ, Kerns SA,
Gargac SM, Pothoulakis AJ. Diastolic augmentation during direct mechanical cardiac actuation. ASAIO J. 2008;54(2):32A.
Anstadt MP, Hendry PJ, Plunkett MD, Menius JA Jr, Pacifico AD Jr, Lowe
JE. Mechanical myocardial actuation during ventricular fibrillation improves tolerance to ischemia compared with cardiopulmonary bypass. Circulation. 1990;
82: IV284-90.
Anstadt MP, Malone JP, Brown GR, Nolan DS, Quinones JD, Anstadt GL.
Direct mechanical ventricular assistance promotes salvage of ischemic myocardium.
ASAIO Trans. 1987;33(3):720-5.
Anstadt MP, Perez-Tamayo RA, Lowe JE, Anstadt GL. Recent Progress Using the Anstadt Cup for Direct Mechanical Ventricular Actuation. In F. Unger (Ed.),
Assisted Circulation.1995; 4:394-408.
Anstadt MP, Stonnington MJ, Tedder M, Crain BJ, Brothers MF, Hilleren DJ,
Rahija RJ, Menius JA Jr, Lowe JE. Pulsatile reperfusion after cardiac arrest improves neurologic outcome. Ann Surg. 1991;214(4), 478-88; discussion 489-90.
Anstadt MP, Taber JE, Hendry PJ, Plunkett MD, Tedder M, Menius JA Jr,
Lowe JE. Myocardial tolerance to ischemia after resuscitation. Direct mechanical ventricular actuation versus cardiopulmonary bypass. ASAIO Trans.1991; 37(3),
M518-9.
Anstadt MP, Tedder M, Hegde SS, Perez-Tamayo RA, Crain BJ, Khian Ha VL,
Abdel-Aleem S, White WD, Lowe JE. Pulsatile versus nonpulsatile reperfusion
175
improves cerebral blood flow after cardiac arrest. Ann Thorac Surg.1993; 56(3),
453-61.
Anstadt MP, Tedder SD, Heide RS, Tedder M, Hilleren DJ, Sostman HD,
Reimer KA, Lowe JE. Cardiac pathology following resuscitative circulatory support. Direct mechanical ventricular actuation versus cardiopulmonary bypass.
ASAIO J.1992;38(2), 75-81.
Artrip JH, Yi GH, Shimizo J, Feihn E, Sciacca RR, Wang J, Burkhoff D.
Maximizing hemodynamic effectiveness of biventricular assistance by direct cardiac compression studied in ex vivo and in vivo canine models of acute heart failure. J
Thorac Cardiovasc Surg. 2000; 120: 379-86.
Baric D. Why pulsatility still matters: a review of current knowledge. Croat Med J.
2014;55(6), 609-20.
Baturalp TB, Ertas A. State of the Art Mock Circulation Loop and a Proposed
Novel Design. Int'l Conf Biomed Eng Sci. 2015; 23-9.
Beaver BV, Reed W, Leary S, McKiernan B, Bain F, Schultz R, Bennett BT,
Pascoe P, Shull E, Cork LC, et al. Report of the AVMA panel on euthanasia. J Am
Vet Med Assoc. 2001; 218: 669-96.
Benjamin EJ, Virani SS, Callaway CW, Chamberlain AM, Chang AR, Cheng
S, Chiuve SE, Cushman M, Delling FN, Deo R, et al. Heart Disease and Stroke
Statistics-2018 Update: A Report From the American Heart Association.
Circulation. 2018;137(12):e67-e492.
176
Bhimaraj A, Uribe C, Suarez EE. Physiological impact of continuous flow on end-organ function: clinical implications in the current era of left ventricular assist devices. Methodist Debakey Cardiovasc J. 2015;11(1):12-7.
Biering-Sørensen T, Biering-Sørensen SR, Olsen FJ, Sengeløv M, Jørgensen
PG, Mogelvang R, Shah AM, Jensen JS. Global Longitudinal Strain by
Echocardiography Predicts Long-Term Risk of Cardiovascular Morbidity and
Mortality in a Low-Risk General Population: The Copenhagen City Heart Study.
Circ Cardiovasc Imaging. 2017;10(3). pii: e005521.
Bircher N, Safar P. Open-chest CPR: an old method whose time has returned. Am
J Emerg Med. 1984;2:568-71.
Blessberger H, Binder T. NON-invasive imaging: Two dimensional speckle tracking echocardiography: basic principles. Heart. 2010;96(9):716-22.
Brown CG, Schlaifer J, Jenkins J, Anstadt GL, Werman HA, Tallman RD Jr,
Ashton J, Hamlin RL, Anstadt MP. Effect of Direct Mechanical Ventricular
Assistance on Myocardial Hemodynamics During Ventricular Fibrillation. Crit Care
Med. 1989;17(11): 1175-80.
Cardiac Arrest Registry to Enhance Survival website. Data reported by ROC and
CARES. https://mycares.net. Accessed April 17, 2017.
Chang WT, Ma MH, Chien KL, Huang CH, Tsai MS, Shih FY, Yuan A, Tsai
KC, Lin FY, Lee YT, et al. Postresuscitation myocardial dysfunction: correlated factors and prognostic implications. Intensive Care Med 2007; 33:88–95.
177
Chelly J, Mongardon N, Dumas F, Varenne O, Spaulding C, Vignaux O, Carli P,
Charpentier J, Pene F, Chiche JD, et al. Benefit of an early and systematic imaging procedure after cardiac arrest: insights from the PROCAT (Parisian Region Out of
Hospital Cardiac Arrest) registry. Resuscitation. 2012;83:1444-1450.
Cheng A, Williamitis CA, Slaughter MS. Comparison of continuous-flow and pulsatile-flow left ventricular assist devices: is there an advantage to pulsatility?
Annals of Cardiothoracic Surgery. 2014;3(6), 573-81.
Cherry BH, Nguyen AQ, Hollrah RA, Olivencia-Yurvati AH, Mallet RT.
Modeling cardiac arrest and resuscitation in the domestic pig. World J Crit Care Med.
2015;4(1):1-12.
Choudhury A, Magoon R, Malik V, Kapoor PM, Ramakrishnan S. Studying diastology with speckle tracking echocardiography: The essentials. Ann Card
Anaesth. 2017;20(Supplement):S57-S60.
Chung M, Shiloh AL, Carlese A. Monitoring of the adult patient on venoarterial extracorporeal membrane oxygenation. Scientific World Journal.
2014;2014:393258.
Cimino S, Canali E, Petronilli V, Cicogna F, De Luca L, Francone M, Sardella
G, Iacoboni C, Agati L. Global and regional longitudinal strain assessed by two-dimensional speckle tracking echocardiography identifies early myocardial dysfunction and transmural extent of myocardial scar in patients with acute ST elevation myocardial infarction and relatively preserved LV function. Eur Heart J
Cardiovasc Imaging. 2013;14(8):805-11.
178
Clark JB, Wang S, Palanzo DA, Wise R, Baer LD, Brehm C, Ündar A. Current
Techniques and Outcomes in Extracorporeal Life Support. Artif Organs.
2015;39(11):926-30.
Cohn LH. Cardiac Surgery in the Adult, 4e. Chapter 17. Cardiopulmonary
Resuscitation. McGraw-Hill Education LLC., 2012. Print.
Crosby JR, DeCook KJ, Tran PL, Betterton E, Smith RG, Larson DF, Khalpey
ZI, Burkhoff D, Slepian MJ. A Physical Heart Failure Simulation System Utilizing the Total Artificial Heart and Modified Donovan Mock Circulation. Artif Organs.
2017;41(7):E52-E65.
Crosby JR, DeCook KJ, Tran PL, Smith RG, Larson DF, Khalpey ZI,
Burkhoff D, Slepian MJ. Physiological Characterization of the SynCardia Total
Artificial Heart in a Mock Circulation System. ASAIO J. 2015;61(3):274-81.
Cummins RO. American Heart Association. ACLS: Principles and Practice. “CPR:
Assessment, Adjuncts, and Alternatives. Circulation. 2003:181-99.
Demirozu ZT, Radovancevic R, Hochman LF, Gregoric ID, Letsou GV, Kar B,
Bogaev RC, Frazier OH. Arteriovenous malformation and gastrointestinal bleeding in patients with the HeartMate II left ventricular assist device. J Heart
Lung Transplant. 2011;30(8):849-53.
Ersbøll M, Andersen MJ, Valeur N, Mogensen UM, Fakhri Y, Thune JJ,
Møller JE, Hassager C, Søgaard P, Køber L. Early diastolic strain rate in relation to systolic and diastolic function and prognosis in acute myocardial infarction: a two-dimensional speckle-tracking study. Eur Heart J. 2014;35(10):648-56.
179
Fischer TG, Schmitt BA, Zhou Y, Reynolds DB, Anstadt MP. Development of
Practical Silicone Ventricles for Testing Direct Mechanical Ventricular Actuation.
Ann Biomed Eng. In: 2015 Biomedical Engineering Society Annual Meeting: 7-10
October 2015; Tampa, FL.
Flachskampf FA, Biering-Sørensen T, Solomon SD, Duvernoy O, Bjerner T,
Smiseth OA. Cardiac Imaging to Evaluate Left Ventricular Diastolic Function.
JACC Cardiovasc Imaging. 2015;8(9):1071-93.
Fukuda N, Terui T, Ishiwata S, Kurihara S. Titin-based regulations of diastolic and systolic functions of mammalian cardiac muscle. Journal of Molecular and
Cellular Cardiology. 2010; 876-81.
Gaffney AM, Wildhirt SM, Griffin MJ, Annich GM, Radomski MW.
Extracorporeal life support. BMJ. 2010 Nov 2;341:c5317.
Geyer H, Caracciolo G, Abe H, Wilansky S, Carerj S, Gentile F, Nesser HJ,
Khandheria B, Narula J, Sengupta PP. Assessment of myocardial mechanics using speckle tracking echocardiography: fundamentals and clinical applications. J
Am Soc Echocardiogr. 2010; 23: 351-69.
Gilotra NA, Stevens GR. Temporary Mechanical Circulatory Support: A Review of the Options, Indications, and Outcomes. Clin Med Insights Cardiol. 2014;
8(Suppl 1): 75–85.
Gräsner JT, Lefering R, Koster RW, Masterson S, Böttiger BW, Herlitz J,
Wnent J, Tjelmeland IB, Ortiz FR, Maurer H, et al. EuReCa ONE—27 Nations,
180
ONE Europe, ONE Registry: a prospective one month analysis of out-of-hospital cardiac arrest outcomes in 27 countries in Europe. Resuscitation. 2016;105:188-95.
Gregory S, Timms D, Pearcy MJ, Tansley G. A naturally shaped silicone ventricle evaluated in a mock circulation loop: a preliminary study. J Med Eng
Technol. 2009;33(3):185-91.
Griffith R, Anstadt MP, Hoekstra J. Regional Cerebral Blood Flow with Manual
Internal Cardiac Massage Versus Direct Mechanical Ventricular Actuation. Ann
Emerg Med. 1992;21(2): 137-41.
Gupta S, and Amanullah AM. Newer modalities for imaging nonischemic cardiomyopathy. Rev Cardiovasc Med. 2015;16(1):51-67.
Hake TG. Studies on ether and chloroform from Prof. Schiff’s physiological laboratory. The Practitioner 1874;12:241-50.
Hallstrom AP, Ornato JP, Weisfeldt M, Travers A, Christenson J, McBurnie
MA, Zalenski R, Becker LB, Schron EB, Proschan M. Public Access
Defibrillation Trial Investigators. Public-access defibrillation and survival after out-of-hospital cardiac arrest. N Engl J Med. 2004;351:637-46.
Halperin HR, Tsitlik JE, Guerci AD, Mellits ED, Levin HR, Shi AY, Chandra
N, Weisfeldt ML. Determinants of blood flow to vital organs during cardiopulmonary resuscitation in dogs. Circulation. 1986;73:539-50.
Hamlin RL, del Rio C. dP/dt(max)--a measure of 'baroinometry'. J Pharmacol
Toxicol Methods. 2012;66(2):63-5.
181
Hashimoto J, Ito S. Pulse pressure amplification, arterial stiffness, and peripheral wave reflection determine pulsatile flow waveform of the femoral artery.
Hypertension. 2010; 56: 926-33.
Horvath MA, Varela CE, Dolan EB, Whyte W, Monahan DS, Payne CJ,
Wamala IA, Vasilyev NV, Pigula FA, Mooney DJ, et al. Towards Alternative
Approaches for Coupling of a Soft Robotic Sleeve to the Heart. Ann Biomed Eng.
2018;46(10):1534-1547.
Huang Y, Gallagher G, Plekhanov S, Morita S, Brady PW, Hunyor SN.
HeartPatch implanted direct cardiac compression: effect on coronary flow and flow patterns in acute heart failure sheep. ASAIO J. 2003;49(3):309-13.
Iribe G, Helmes M, Kohl P. Force-length relations in isolated intact cardiomyocytes subjected to dynamic changes in mechanical load. American
Journal of Physiology: Heart and Circulatory Physiology. 2007; 292: H1487-97.
Itoh H, Ichiba S, Ujike Y, Douguchi T, Obata H, Inamori S, Iwasaki T,
Kasahara S, Sano S, Ündar A. Effect of the Pulsatile Extracorporeal Membrane
Oxygenation on Hemodynamic Energy and Systemic Microcirculation in a Piglet
Model of Acute Cardiac Failure. Artif Organs. 2016;40(1):19-26.
Jalife J. Ventricular fibrillation: mechanisms of initiation and maintenance. Annu
Rev Physiol. 2000;62:25-50.
Jentzer JC, Anavekar NS, Mankad SV, Khasawneh M, White RD, Barsness
GW, Rabinstein AA, Kashani KB, Pislaru SV. Echocardiographic left ventricular
182
diastolic dysfunction predicts hospital mortality after out-of-hospital cardiac arrest. J
Crit Care. 2018;47:114-120.
Jentzer JC, Anavekar NS, Mankad SV, White RD, Kashani KB, Barsness GW,
Rabinstein AA, Pislaru SV. Changes in left ventricular systolic and diastolic function on serial echocardiography after out-of-hospital cardiac arrest.
Resuscitation. 2018;126:1-6.
Jentzer JC, Chonde MD, Dezfulian C. Myocardial Dysfunction and Shock after
Cardiac Arrest. Biomed Res Int. 2015; 2015: 314796.
Jiang L, Zhang JS. Mechanical cardiopulmonary resuscitation for patients with cardiac arrest. World J Emerg Med. 2011;2(3):165-8.
Kaebnick BW, Giridharan GA, Koenig SC. Quantification of pulsatility as a function of vascular input impedance: an in vitro study. ASAIO J.
2007;53(2):115-21.
Kato TS, Chokshi A, Singh P, Khawaja T, Cheema F, Akashi H, Shahzad K,
Iwata S, Homma S, Takayama H, et al. Effects of continuous-flow versus pulsatile-flow left ventricular assist devices on myocardial unloading and remodeling. Circ Heart Fail. 2011; 4: 546-53.
Kawahito S, Takano T, Nakata K, Maeda T, Nonaka K, Linneweber J,
Schulte-Eistrup S, Sato T, Mikami M, Glueck J, et al. Analysis of the arterial blood pressure waveform during left ventricular nonpulsatile assistance in animal models. Artif Organs. 2000; 24: 816-20.
183
Kern KB, Hilwig RW, Berg RA, Rhee KH, Sanders AB, Otto CW, Ewy GA.
Postresuscitation left ventricular systolic and diastolic dysfunction. Treatment with dobutamine. Circulation. 1997;95(12):2610-3.
Kho C, Lee A, Jeong D, Oh JG, Chaanine AH, Kizana E, Park WJ, Hajjar RJ.
SUMO1-dependent modulation of SERCA2a in heart failure. Nature. 2011; 477:
601-5.
Kresh JY, Chopra A. Intercellular and extracellular mechanotransduction in cardiac myocytes. Pflugers Arch. 2011; 462:75-87.
Krüger M, Linke WA. Titin-based mechanical signaling in normal and failing myocardium. J Mol Cell Cardiol. 2009;46(4):490-8.
Kung RT, Rosenberg M. Heart booster: a pericardial support device. Ann Thorac
Surg. 1999;68(2):764-7.
Lang ES, Al Raisi M. Ventricular tachyarrhythmias (out-of-hospital cardiac arrests). BMJ Clin Evid;2006.pii: 0216.
Lee J; National Clinical Trial Initiative Subcommittee. Long-term Mechanical
Circulatory Support System reliability recommendation by the National Clinical
Trial Initiative subcommittee. ASAIO J. 2009;55(6):534-42.
Leung DY, Ng AC. Emerging clinical role of strain imaging in echocardiography.
Heart Lung Circ. 2010;19(3):161-74.
Li BN, Dong MC, Vai MI. On an automatic delineator for arterial blood pressure waveforms. Biomedical Signal Processing and Control. 2010; 5: 76-81.
184
Liang YJ, Zhang Q, Fang F, Lee AP, Liu M, Yan BP, Lam YY, Chan GC, Yu
CM. Incremental value of global systolic dyssynchrony in determining the occurrence of functional mitral regurgitation in patients with left ventricular systolic dysfunction. Eur Heart J. 2013;34(10):767-74.
Lim P, Buakhamsri A, Popovic ZB, Greenberg NL, Patel D, Thomas JD,
Grimm RA. Longitudinal strain delay index by speckle tracking imaging: a new marker of response to cardiac resynchronization therapy. Circulation.
2008;118(11):1130-7.
Link MS, Berkow LC, Kudenchuk PJ, Halperin HR, Hess EP, Moitra VK,
Neumar RW, O'Neil BJ, Paxton JH, Silvers SM, et al. Part 7: Adult Advanced
Cardiovascular Life Support: 2015 American Heart Association Guidelines Update for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care.
Circulation. 2015;132(18 Suppl 2):S444-64.
Lowe JE, Anstadt MP, Van Trigt P, Smith PK, Hendry PJ, Plunkett MD,
Anstadt GL. First successful bridge to cardiac transplantation using direct mechanical ventricular actuation. Ann Thorac Surg. 1991;52(6):1237-43; discussion
1243-5.
Lowe JE, Hughes G, Biswas S. Non-Blood-Contacting Biventricular Support:
Direct Mechanical Ventricular Actuation. Operative Techniques in Thoracic and
Cardiovascular Surgery. 1999;4(4):345-51.
Mandawat A; Rao SV. Percutaneous Mechanical Circulatory Support Devices in
Cardiogenic Shock. Circ Cardiovasc Interv. 2017;10(5). pii: e004337.
185
McCain ML, Parker KK. Mechanotransduction: the role of mechanical stress, myocyte shape, and cytoskeletal architecture on cardiac function. Pflugers Arch.
2011; 462: 89-104.
McConnell PI, Anstadt MP, Del Rio CL, Preston TJ, Ueyama Y, Youngblood
BL. Cardiac function after acute support with direct mechanical ventricular actuation in chronic heart failure. ASAIO J. 2014;60(6):701-6.
Merchant RM, Yang L, Becker LB, Berg RA, Nadkarni V, Nichol G, Carr BG,
Mitra N, Bradley SM, Abella BS, et al. Incidence of treated cardiac arrest in hospitalized patients in the United States. Crit Care Med. 2011;39(11):2401-6.
Miller PE, Solomon MA, McAreavey D. Advanced Percutaneous Mechanical
Circulatory Support Devices for Cardiogenic Shock. Crit Care Med.
2017;45(11):1922-29.
Mor-Avi V, Lang RM, Badano LP, Belohlavek M, Cardim NM, Derumeaux G,
Galderisi M, Marwick T, Nagueh SF, Sengupta PP, et al. Current and evolving echocardiographic techniques for the quantitative evaluation of cardiac mechanics:
ASE/EAE consensus statement on methodology and indications endorsed by the
Japanese Society of Echocardiography. J Am Soc Echocardiogr.
2011;24(3):277-313.
Moreno MR, Biswas S, Harrison LD, Pernelle G, Miller MW, Fossum TW,
Nelson DA, Criscione JC. Development of a Non-Blood Contacting Cardiac Assist and Support Device: An In Vivo Proof of Concept Study. J Med Devices.
2011;5(4):041007-1-9.
186
Moreno MR, Biswas S, Harrison LD, Pernelle G, Miller MW, Fossum TW,
Nelson DA, Criscione JC. Assessment of Minimally Invasive Device That
Provides Simultaneous Adjustable Cardiac Support and Active Synchronous Assist in an Acute Heart Failure Model. 2011;5(4):041008-1-9.
Morrissey C. Echo for diastology. Ann Card Anaesth. 2016;19
(Supplement):S12-S18.
Mozaffarian D, Benjamin EJ, Go AS, Arnett DK, Blaha MJ, Cushman M, Das
SR, de Ferranti S, Després JP, Fullerton HJ, et al. Heart Disease and Stroke
Statistics-2016 Update: A Report From the American Heart Association. Circulation.
2016;133(4):e38-60.
Nagpal AD, Singal RK, Arora RC, Lamarche Y. Temporary Mechanical
Circulatory Support in Cardiac Critical Care: A State of the Art Review and
Algorithm for Device Selection. Can J Cardiol. 2017;33(1):110-8.
Nagueh SF, Smiseth OA, Appleton CP3, Byrd BF 3rd, Dokainish H, Edvardsen
T, Flachskampf FA, Gillebert TC, Klein AL, Lancellotti P, Marino P, et al.
Recommendations for the Evaluation of Left Ventricular Diastolic Function by
Echocardiography: An Update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. J Am Soc Echocardiogr.
2016;29(4):277-314.
Neumar RW, Nolan JP, Adrie C, Aibiki M, Berg RA, Böttiger BW, Callaway
C, Clark RS, Geocadin RG, Jauch EC, et al. Post-cardiac arrest syndrome: epidemiology, pathophysiology, treatment, and prognostication. A consensus
187
statement from the International Liaison Committee on Resuscitation. Circulation.
2008;118(23):2452-83.
Nichol G, Thomas E, Callaway CW, Hedges J, Powell JL, Aufderheide TP, Rea
T, Lowe R, Brown T, Dreyer J, et al. Regional variation in out-of-hospital cardiac arrest incidence and outcome. JAMA. 2008;300(12):1423-31.
O'Neil MP, Fleming JC, Badhwar A, Guo LR. Pulsatile versus nonpulsatile flow during cardiopulmonary bypass: microcirculatory and systemic effects. Ann Thorac
Surg. 2012;94(6):2046-53.
Ouweneel DM, Henriques JPS. Percutaneous cardiac support devices for cardiogenic shock: current indications and recommendations. Heart. 2012;
98:1246-54.
Oz MC, Artrip JH, Burkhoff D. Direct cardiac compression devices. J Heart Lung
Transplant. 2002;21(10):1049-55.
Pandit SV, Jalife J. Rotors and the dynamics of cardiac fibrillation. Circ Res.
2013;112(5):849-62.
Pantalos GM, Altieri F, Berson A, Borovetz H, Butler K, Byrd G, Ciarkowski
AA, Dunn R, Frazier OH, Griffith B, et al. Long-term mechanical circulatory support system reliability recommendation: American Society for Artificial Internal
Organs and The Society of Thoracic Surgeons: long-term mechanical circulatory support system reliability recommendation. Ann Thorac Surg. 1998;66(5):1852-9.
Pantridge JF, Geddes JS. Cardiac arrest after myocardial infarction. Lancet.
1966;1:807-8.
188
Papaioannou TG, Mathioulakis DS, Tsangaris SG. Simulation of systolic and diastolic left ventricular dysfunction in a mock circulation: the effect of arterial compliance. J Med Eng Technol. 2003;27(2):85-9.
Park JJ, Park JB, Park JH, Cho GY. Global Longitudinal Strain to Predict
Mortality in Patients with Acute Heart Failure. J Am Coll Cardiol.
2018;71(18):1947-57.
Patil KD, Halperin HR, Becker LB. Cardiac arrest: resuscitation and reperfusion.
Circ Res. 2015;116(12):2041-9.
Perez-Tamayo RA, Anstadt MP, Cothran RL Jr, Reisinger RJ, Schenkman DI,
Hulette C, Reimer KA, Anstadt GL, Lowe JE. Prolonged total circulatory support using direct mechanical ventricular actuation. ASAIO J. 1995;41(3):M512-7.
Prosser BL, Khairallah RJ, Ziman AP, Ward CW, Lederer WJ. X-ROS signaling in the heart and skeletal muscle: Stretch-dependent local ROS regulates
2+ [Ca ]i. J Mol Cell Cardiol. 2013; 58: 172-81.
Prosser BL, Ward CW, Lederer WJ. X-ROS signaling: rapid mechano-chemo transduction in heart. Science. 2011; 333: 1440-5.
Prosser BL, Ward CW. Mechano-chemo transduction tunes the heartstrings. Sci
Signal. 2014;7(317):pe7.
Roche ET, Horvath MA, Wamala I, Alazmani A, Song SE, Whyte W,
Machaidze Z, Payne CJ, Weaver JC, Fishbein G, et al. Soft robotic sleeve supports heart function. Sci Transl Med. 2017;9(373).
189
Romagnoli S, Romano SM, Bevilacqua S, Lazzeri C, Gensini GF, Pratesi C,
Quattrone D, Dini D, De Gaudio AR. Dynamic response of liquid-filled catheter systems for measurement of blood pressure: precision of measurements and reliability of the Pressure Recording Analytical Method with different disposable systems. J Crit Care. 2011; 26: 415-22.
Schmitt BA, Wright NV, Zhou Y, Reynolds DB, Anstadt MP. Vascular
Pulsatility During Mechanical Actuation of the Fibrillating Heart is Equivalent to the Native Beating Heart. Circulation. In: 2016 American Heart Association
Scientific Sessions: 12-16 November 2016; New Orleans, LA.
Schmitt BA, Zhou Y, Wright NV, Reynolds DB, Anstadt MP. Pulsatile
Characteristics of the Mechanically Actuated Fibrillating Heart are Similar to the
Native Beating Heart. Ann Thorac Surg. In: 2017 Society of Thoracic Surgeon
Annual Meeting: 21-25 January 2017; Houston, TX.
Segers P, Rietzschel ER, De Buyzere ML, Stergiopulos N, Westerhof N, Van
Bortel LM, Gillebert T, Verdonck PR. Three- and four-element Windkessel models: assessment of their fitting performance in a large cohort of healthy middle-aged individuals. Proc Inst Mech Eng H. 2008;222(4):417-28.
Sekhon MS, Ainslie PN, Griesdale DE. Clinical pathophysiology of hypoxic ischemic brain injury after cardiac arrest: a “two-hit” model. Crit Care. 2017; 21:
90.
190
Sen A, Larson JS, Kashani KB, Libricz SL, Patel BM, Guru PK, Alwardt CM,
Pajaro O, Farmer JC. Mechanical circulatory assist devices: a primer for critical care and emergency physicians. Crit Care. 2016;20(1):153.
Sénage T, Février D, Michel M, Pichot E, Duveau D, Tsui S, Trochu JN,
Roussel JC. A mock circulatory system to assess the performance of continuous-flow left ventricular assist devices (LVADs): does axial flow unload better than centrifugal LVAD? ASAIO J. 2014;60(2):140-7.
Sharov VS, Dremina ES, Galeva NA, Williams TD, Schöneich C. Quantitative mapping of oxidation-sensitive cysteine residues in SERCA in vivo and in vitro by
HPLC-electrospray-tandem MS: selective protein oxidation during biological aging.
Biochem J. 2006; 394: 605-15.
Shepard RB, Simpson DC, Sharp JF. Energy equivalent pressure. Arch Surg.
1966;93: 730-40.
Siao FY, Chiu CC, Chiu CW, Chen YC, Chen YL, Hsieh YK, Lee CH, Wu CT,
Chou CC, Yen HH. Managing cardiac arrest with refractory ventricular fibrillation in the emergency department: Conventional cardiopulmonary resuscitation versus extracorporeal cardiopulmonary resuscitation. Resuscitation.
2015;92:70-6.
Smiseth OA, Torp H, Opdahl A, Haugaa KH, Urheim S. Myocardial strain imaging: how useful is it in clinical decision making? Eur Heart J.
2016;37(15):1196-207.
191
Soucy KG, Koenig SC, Giridharan GA, Sobieski MA, Slaughter MS. Defining pulsatility during continuous-flow ventricular assist device support. J Heart Lung
Transplant. 2013;32(6):581-7.
Soucy KG, Koenig SC, Giridharan GA, Sobieski MA, Slaughter MS. Rotary pumps and diminished pulsatility: do we need a pulse? ASAIO J. 2013; 59: 355-66.
Stub D, Bernard S, Pellegrino V, Smith K, Walker T, Sheldrake J, Hockings L,
Shaw J, Duffy SJ, Burrel A, et al. Refractory cardiac arrest treated with mechanical CPR, hypothermia, ECMO and early reperfusion (the CHEER trial).
Resuscitation. 2015;86:88-94.
Suarez J, Patel CB, Felker GM, Becker R, Hernandez AF, Rogers JG.
Mechanisms of bleeding and approach to patients with axial-flow left ventricular assist devices. Circ Heart Fail. 2011;4(6):779-84.
Timms D, Hayne M, McNeil K, Galbraith A. A complete mock circulation loop for the evaluation of left, right, and biventricular assist devices. Artif Organs.
2005;29(7):564-72.
Timms DL, Gregory SD, Greatrex NA, Pearcy MJ, Fraser JF, Steinseifer U. A compact mock circulation loop for the in vitro testing of cardiovascular devices.
Artif Organs. 2011;35(4):384-91.
Tomaselli GF. Introduction to a compendium on sudden cardiac death: epidemiology, mechanisms, and management. Circ Res. 2015;116(12):1883-6.
Travis AR, Giridharan GA, Pantalos GM, Dowling RD, Prabhu SD, Slaughter
MS, Sobieski M, Undar A, Farrar DJ, Koenig SC. Vascular pulsatility in patients
192
with a pulsatile- or continuous-flow ventricular assist device. J Thorac Cardiovasc
Surg. 2007;133(2):517-24.
Undar A, Fraser CD Jr. The alphabet of research on pulsatile and nonpulsatile
(continuous flow) perfusion during chronic support. Artif Organs. 2002;26(9):812-3; author reply 813-5.
Undar A, Frazier OH, Fraser CD Jr. Defining pulsatile perfusion: quantification in terms of energy equivalent pressure. Artif Organs. 1999; 23: 712-16.
Undar A, Zapanta CM, Reibson JD, Souba M, Lukic B, Weiss WJ, Snyder AJ,
Kunselman AR, Pierce WS, Rosenberg G, et al. Precise quantification of pressure flow waveforms of a pulsatile ventricular assist device. ASAIO J. 2005; 51: 56-9.
Undar A. Pulsatile versus nonpulsatile cardiopulmonary bypass procedures in neonates and infants: from bench to clinical practice. ASAIO J. 2005;51(5):vi-x.
Vognsen M, Fabian-Jessing BK, Secher N, Løfgren B, Dezfulian C, Andersen
LW, Granfeldt A. Contemporary animal models of cardiac arrest: A systematic review. Resuscitation. 2017;113:115-23.
Voigt J, Pedrizzetti G, Lysyansky P, Marwick TH, Houle H, Baumann R, Pedri
S, Ito Y, Abe Y, Metz S, et al. Definitions for a common standard for 2D speckle tracking echocardiography: consensus document of the EACVI/ASE/Industry Task
Force to standardize deformation imaging. J Am Soc Echocardiogr. 2015; 28:
183-93.
193
Wang J, Khoury DS, Thohan V, Torre-Amione G, Nagueh SF. Global Diastolic
Strain Rate for the Assessment of Left Ventricular Relaxation and Filling Pressures.
Circulation. 2007;115(11):1376-83.
Weisfeldt ML, Becker LB. Resuscitation after cardiac arrest: a 3-phase time-sensitive model. JAMA. 2002;288:3035-8.
Westerhof N, Lankhaar JW, Westerhof BE. The arterial Windkessel. Med Biol
Eng Comput. 2009;47(2):131-41.
Wever-Pinzon O, Selzman CH, Drakos SG, Saidi A, Stoddard GJ, Gilbert EM,
Labedi M, Reid BB, Davis ES, Kfoury AG, et al. Pulsatility and the Risk of
Nonsurgical Bleeding in Patients Supported With the Continuous-Flow Left
Ventricular Assist Device HeartMate II. Circ Heart Fail. 2013;6(3):517-26.
Wozniak CJ, Pothoulakis AJ, Gargac SM, Darner RJ, Anstadt MP.
Echocardiographic assessment of direct mechanical ventricular actuation. ASAIO.
2007;53(2):40A.
Wozniak CJ, Pothoulakis AJ, Gargac SM, Darner RJ, Anstadt MP. Velocity vector imaging characterizes systolic and diastolic augmentation of the failing heart during non-blood contacting ventricular actuation. JACC. 2007;49(9):s1, 70A.
Xu T, Tang W, Ristagno G, Wang H, Sun S, Weil MH. Postresuscitation myocardial diastolic dysfunction following prolonged ventricular fibrillation and cardiopulmonary resuscitation. Crit Care Med. 2008;36(1):188-92.
Yousuf O, Chrispin J, Tomaselli GF, Berger RD. Clinical management and prevention of sudden cardiac death. Circ Res. 2015;116(12):2020-40.
194
Yu CM, Chau E, Sanderson JE, Fan K, Tang MO, Fung WH, Lin H, Kong SL,
Lam YM, Hill MR, et al. Tissue Doppler echocardiographic evidence of reverse remodeling and improved synchronicity by simultaneously delaying regional contraction after biventricular pacing therapy in heart failure. Circulation.
2002;105(4):438-45.
Yu CM, Fung JW, Zhang Q, Chan CK, Chan YS, Lin H, Kum LC, Kong SL,
Zhang Y, Sanderson JE. Tissue Doppler imaging is superior to strain rate imaging and postsystolic shortening on the prediction of reverse remodeling in both ischemic and nonischemic heart failure after cardiac resynchronization therapy. Circulation.
2004;110(1):66-73.
Zhou Y, Schmitt BA, Anstadt MP. Left Ventricular Diastolic Function Is
Returned during Direct Mechanical Ventricular Actuation of the Arrested Heart.
Circulation. In: 2018 American Heart Association Annual Meeting Scientific
Sessions 2018: 10-12 November 2018; Chicago, IL.
Zhou Y, Schmitt BA, Anstadt MP. Mechanical Ventricular Actuation during
Ventricular Fibrillation Results in Near-Physiological Left Ventricular Myocardial
Mechanics. Circulation. In: 2018 American Heart Association Annual Meeting
Scientific Sessions 2018: 10-12 November 2018; Chicago, IL.
Zhou Y, Schmitt BA, Reynolds DB, Anstadt MP. Echocardiographic
Characterization of a Biventricular Mock Circulatory System and a Fibrillating
Heart during Mechanical Actuation. J Am Soc Echocardiogr. In: 2017 American
Society of Echocardiography Annual Meeting: 2-6 June 2017; Baltimore, MD.
195
Zhou Y, Schmitt BA, Reynolds DB, Anstadt MP. Use of a Mock Ventricle to
Simulate Functionality of the Fibrillating Heart for Testing Direct Cardiac
Compression Devices. ASAIO J. In: 2017 American Society for Artificial Internal
Organs Annual Meeting: 21-24 June 2017; Chicago, IL.
196
APPENDIX A Commonly Used Abbreviations
ACLS Advanced Cardiac Life Support
AED Automatic External Defibrillators
AHA American Heart Association
AIDS Acquired Immunodeficiency Syndrome
AV Atrioventricular
BTD Bridge to Decision
BTT Bridge to Transplantation
CA Cardiac Arrest
CC Close Chest
CO Cardiac Output
CPB Cardiopulmonary Bypass
CPI Cardiac Power Integral
CPR Cardiopulmonary Resuscitation
CRT Cardiac Resynchronization Therapy
DCC Direct Cardiac Compression dGLS Diastolic Global Longitudinal Strain dGLSR Diastolic Global Longitudinal Strain Rate
DI Dyssynchrony Index
DMVA Direct Mechanical Ventricular Actuation dRLS Diastolic Regional Longitudinal Strain dRLSR Diastolic Regional Longitudinal Strain Rate
197
DT Destination Therapy ECG Electrocardiogram ECLS Extracorporeal Life Support
ECMO Extracorporeal Membrane Oxygenation
EDPVR End-Diastolic Pressure-Volume Relationship
EDV End-Diastolic Volume
EDVI End-Diastolic Volume Index
EEP Energy Equivalent Pressure
EF Ejection Fraction
ESPVR End-Systolic Pressure-Volume Relationship
ESV End-Systolic Volume
ESVI End-Systolic Volume Index
GLS Global Longitudinal Strain
GLSR Global Longitudinal Strain Rate
GWTG Get With The Guidelines
HF Heart Failure
ICD Implantable Cardioverter Defibrillator
IHCA In-Hospital Cardiac Arrest
LV Left Ventricle LVAD Left Ventricular Assist Device
LVEDP Left Ventricle End-Diastolic Pressure
LVESP Left Ventricle End-Systolic Pressure
MAP Mean Arterial Pressure
MCL Mock Circulation Loop
MCS Mechanical Circulatory Support NBH Native Beating Heart
OC Open Chest
OHCA Out-of-Hospital Cardiac Arrest
198
PAMD Post-arrest Myocardial Dysfunction
PI Pulsatility Index
PP Pulse Pressure
PPI Pulse Power Index pVAD Percutaneous Ventricular Assist Device
QOL Quality of Life
RLS Regional Longitudinal Strain
RLSR Regional Longitudinal Strain Rate
ROC Resuscitation Outcomes Consortium
ROSC Return of Spontaneous Circulation
RV Right Ventricle SA Specific Aim
SCA Sudden Cardiac Arrest SEM Standard Error of the Mean SERCA Sarcoplasmic/Endoplasmic Reticulum Calcium ATPase sGLS Systolic Global Longitudinal Strain sGLSR Systolic Global Longitudinal Strain Rate
SHE Surplus Hemodynamic Energy sRLS Systolic Regional Longitudinal Strain sRLSR Systolic Regional Longitudinal Strain Rate
STE Speckle Tracking Echocardiography
SV Stroke Volume SVI Stroke Volume Index
SW Stroke Work
TDI Tissue Doppler Imaging
VA Venous-Arterial
VAD Ventricular Assist Device
VF Ventricular Fibrillation
199
VT Ventricular Tachycardia
VV Venous-Venous VVI Velocity Vector Imaging 2-D Two Dimensional 3-D Three Dimensional
200