Anatomical, Biomechanical, and End-Of-Life Considerations for Emergent Cardiac Pacing Technologies

Anatomical, Biomechanical, and End-of-Life Considerations for Emergent Cardiac Pacing Technologies

A DISSERTATION SUBMITTED TO THE FACULTY OF UNIVERSITY OF MINNESOTA BY

ALEXANDER R MATTSON

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY IN BIOMEDICAL ENGINEERING

ADVISOR: PAUL A. IAIZZO, PHD

July 2018

Acknowledgements

My most sincere thanks to the PhD students whose passion, dedication, and friendship defined my graduate life: Lars Mattison, Megan Schmidt, Erik

Gaasedelen, Mikayle Holm, Jorge Zhingre-Sanchez, Brian Howard, Steve

Quallich, David Ramirez. I am beyond fortunate to have worked with such a talented group of people.

Thanks to the lab staff, scientists, and Medtronic collaborators who’ve provided countless hours of work on my behalf: Mike Eggen, Monica Mahre, Tinen Iles,

Alex Deakyne, Mike Bateman, Alison Weyer, Charles Soule, Weston Upchurch.

A special thanks to Paul Iaizzo for taking a chance on the quiet student. Your mentorship was and is invaluable. You’ve helped me put the tools in my toolbox to become the best scientist I can be.

Dedication

This thesis is dedicated to my parents: Dave and Gloria Mattson. I could write a blurb about how much you’ve done for me, but it wouldn’t say nearly enough.

Thank you for everything.

Abstract

Over 600,000 permanent pacing systems are implanted each calendar year as the primary therapy for symptomatic bradycardia. Innovations in pacing technology have rapidly expanded the indications for this life-saving therapy, while reducing complication rates. This thesis examined three prongs of emergent pacing technologies: leadless pacing, epicardial/extravascular pacing, and physiologic pacing through the bundle of His. First, I quantitatively evaluated the likely target anatomies for next-generation pacing systems. Then, anatomic data was supplemented with biomechanics, to provide the foundation upon which next-generation leadless pacemaker fixation mechanisms may be built. Finally, I investigated some of the challenges of extracting leadless pacing systems. The data in this thesis provided a substrate for the design and implementation of next-generation pacing systems.

iii

Table of Contents

List of Tables………………………………………………………………………...…vii List of Figures ………………………………………..………….…………………..viii Section I. An Introduction to Cardiac Pacemakers: Where have we been? Where are we now? What’s next on the Horizon? ...... 1 Chapter 1: The Cardiac Pacemaker: A Crossroads of Engineering and Medicine 2 1.1 Preface ...... 3 1.2 Looking back: A brief history of cardiac pacing ...... 4 1.3 Lillehei and Zoll: World War II Innovators ...... 5 1.4 Earl Bakken and the Birth of a Medical Device Industry ...... 9 1.5 Innovation Fuels Innovation ...... 11 1.6 Modern Implantable Pacing Systems ...... 16 1.7 Implantable Pulse Generators (IPGs) ...... 18 1.8 Sensing Algorithms...... 19 1.9 Pacemaker Leads ...... 20 1.10 Clinical Pacing Categories ...... 24 1.11 Future Pacing Directions ...... 30 1.12 Conclusion ...... 42 Section II. Anatomic Considerations for Next-Generation Pacemakers ...... 43 Chapter 2: The Quantitative Assessment of Epicardial Fat Distributions on Human Hearts: Implications for Epicardial Electrophysiology ...... 45 2.1 Preface ...... 46 2.2 Synopsis ...... 47 2.3 Introduction ...... 49 2.4 Objective ...... 51 2.5 Methods ...... 51 2.6 Results ...... 54 2.7 Discussion ...... 58 2.8 Conclusion ...... 61 Chapter 3: 3-Dimensional Reconstruction and Quantitative Assessment of Ventricular Epicardial Fat Tissue on Human Hearts ...... 63 3.1 Preface ...... 64 3.2 Synopsis ...... 65 3.3 Introduction ...... 67 3.4 Methods ...... 69 3.5 Results ...... 73 3.6 Discussion ...... 82 3.7 Conclusion ...... 87 Chapter 4: Electrical Parameters of Physiologic His-Purkinje Pacing Vary by Implant Location in a Canine Model ...... 88 4.1 Preface ...... 89

4.2 Synopsis ...... 90 4.3 Introduction ...... 92 4.4 Methods ...... 93 4.5 Results ...... 95 4.6 Discussion ...... 101 4.7 Conclusion ...... 109 Chapter 5: Visible Heart Visualization of Physiologic His-Bundle Pacing and Surrounding Anatomy within Reanimated Human Hearts ...... 110 5.1 Preface ...... 111 5.2 Synopsis ...... 111 5.3 Introduction ...... 113 5.4 Case Report ...... 114 5.5 Discussion and Conclusions ...... 119 Chapter 6: 3-Dimensional Anatomic Assessment of the Human Right Atrial Appendage: Implications for Atrial Fixation Mechanisms ...... 122 6.1 Preface ...... 123 6.2 Synopsis ...... 123 6.3 Background ...... 125 6.4 Methods ...... 126 6.5 Results ...... 130 6.6 Discussion ...... 135 6.7 Conclusions ...... 136 Section III. Biomechanics Informing Fixation of Next-Generation Pacing Systems 137 Chapter 7: Perforation Properties of the Right Atrial Appendage ...... 138 7.1 Preface ...... 139 7.2 Synopsis ...... 140 7.3 Introduction and Background ...... 142 7.4 Methods ...... 143 7.5 Results ...... 144 7.6 Discussion ...... 148 7.7 Conclusion ...... 149 Chapter 8: Assessing the Relationship between Right Atrial Stiffness and Chamber Pressure to Quantitatively Define Myocardial Tensile Properties .. 150 8.1 Preface ...... 151 8.2 Synopsis ...... 152 8.3 Background ...... 154 8.4 Methods ...... 155 8.5 Results ...... 158 8.6 Discussion ...... 160 Section IV. End-of-Life Consideration for Leadless Pacemakers ...... 161 Chapter 9: Retrieval of a Chronically Implanted Leadless Pacemaker Within an Isolated Heart using Direct Visualization ...... 162 9.1 Preface ...... 163

9.2 Synopsis ...... 164 9.3 Introduction ...... 166 9.4 Case Report ...... 166 9.5 Discussion and Conclusions ...... 168 Chapter 10: Atraumatic Extrication of Leadless Pacemaker Tines from the Tricuspid Valve Apparatus ...... 171 10.1 Preface ...... 172 10.2 Synopsis ...... 172 10.3 Introduction ...... 175 10.4 Objective ...... 177 10.5 Methods ...... 177 10.6 Results ...... 180 10.7 Discussion ...... 184 10.8 Conclusion ...... 186 Thesis Summary ...... 187 Section V. Bibliography ...... 188

List of Tables Table 2.1: Donor Patient Demographics/Characteristics and Relation to Adipose Coverage ...... 55

Table 3.2: Donor Patient Histories for Hearts Depicted in Figure 3.3-3.5 ...... 73

Table 3.1: Donor Patient Demographics/Characteristics and Relation to Overall Mean RV and LV Adipose Thickness ...... 81

Table 10.1: Patient histories for each donated specimen. Full patient histories, reference by heart number, are available at the University of Minnesota’s Atlas of Human Cardiac Anatomy free access website (vhlab.umn.edu/atlas/histories/histories.shtml)...... 179

Table 10.2: Medical Histories for Donated Specimens ...... 179

vii

List of Figures Figure 1.1 Albert Hyman's original device. This "pacemaker" was designed to provide electrical shocks at the turn of a hand crank, revitalizing a stopped heart...... 5

Figure 1.2: C Walton Lillehei performs open-heart surgery on a patient at the University of Minnesota. Lillehei pioneered many open-heart surgeries and played a pivotal role in bringing pacemakers to mainstream clinical practice...... 7

Figure 1.3: A patient outfitted with the Zoll pacemaker. Ambulatory movement required unplugging and replugging the pulse generator as it moved through the corridor. Incidentally, this patient was the first to receive a transvenous pacing lead implanted by Dr. Seymour Furman...... 8

Figure 1.4: Earl Bakken's first transistorized pacemaker was developed over a course of weeks (A), animal tested over a course of hours (B), before first-in- human implantation...... 10

Figure 1.5: A patient wearing a Medtronic model 5800 pacemaker. The first commercialized pacemakers were worn around the neck. There were common clinical complications with the lead entry sites...... 12

Figure 1.6: The Chardack Greatbatch Pacemaker was licensed by Medtronic in 1960, and quickly became the standard-bearer for implantable pacing devices. The pacemaker can was roughly the size of a hockey puck. Coupled with the Norman-Roth electrode, the Chardack-Greatbatch pacemaker was the benchmark pacing technology throughout the 1960s...... 13

Figure 1.7: Pacing electrodes invented by Dr. Samuel Hunter and Medtronic engineer Hunter Roth. The stainless steel Hunter-Roth electrode helped to reduce pacing capture threshold...... 14

Figure 1.8: Dr. Samuel Hunter (inset) and Medtronic engineer Norman Roth developed a bipolar electrode that represented a major advance in pacing technology. First implanted in 1959, the Hunter-Roth lead helped contribute 7 years of life to a Stokes-Adams disease patient, Warren Mauston in 1960...... 14

Figure 1.9: Pulse generators have dramatically reduced in size over the years. Pictured here is the Medtronic 5800 (the first commercialized transistorized pacemaker), the Medtronic 5950 (implanted in the late 1970s), and the Medtronic Adapta (currently market-available)...... 19

Figure 1.10: Examples of the distal portion of multiple cardiac pacing leads. Pictured here from top to bottom: the Medtronic 4968 epicardial pacing lead, the Medtronic 4074 passive fixation pacing lead, the Medtronic 5076 active fixation pacing lead...... 21 viii

Figure 1.11: Pacing leads come equipped with either passive or active fixation. Shown here are a tined passive fixation and a helical active fixation (cork-screw) mechanism...... 23

Figure 1.12: A cardiac resynchronization therapy pacing and defibrillation system is shown, with the device can implanted in the left pectoral region. Endocardial leads pace within the right atrial and right ventricular chambers of the heart. The defibrillation lead within the right ventricle allows for defibrillation shocks in the case of abnormal ventricular tachyarrhythmia The left ventricular lead enters the coronary sinus, and wraps around the coronary venous system to pace on the left ventricular epicardium...... 27

Figure 1.13: A--The cardiac conduction system is highlighted in red. The bundle of His penetrates through the central fibrous body separating the atrial and ventricular halves of the heart. Shortly after, the bundle of His bifurcates into the right and left bundle branches. B—A Medtronic 3830 lead (Medtronic,, Minneapolis, MN) is shown fixated to the bundle of His in a reanimated swine heart...... 29

Figure 1.14: Image of a Medtronic MicraTM implanted within the right ventricular apex of a reanimated human heart at the University of Minnesota's Visible Heart Laboratory...... 32

Figure 1.15: Encapsulation on the body of chronically implanted leadless pacemakers varies. Devices may be fully encapsulated within a year of implantation, or completely bare several years post-implant. Image borrowed from Reddy et al.38...... 35

Figure 1.16: The WiSE-CRT endocardial LV pacing system. The system utilizes a co-implanted right-sided pacemaker, an ultrasonic transducer, and an endocardial leadless receiver...... 38

Figure 1.17: Attain Performa quadripolar lead family. Courtesy of Medtronic. .... 41

Figure 2.1: 3-Dimensional reconstruction depicting the epicardial adipose coverage of one donated specimen (HH099, 46 year old male with history of CAD). Images of each donated specimen were taken in three anatomical planes: A--Anterior Right Ventricle, B--Left Ventricular Margin, C--Diaphragmatic Surface...... 52

Figure 2.2: A—Hearts were imaged in isolated planes, with appropriately positioned scales so to reference dimensions. Regions with and without adipose both were quantified. Common anatomical points selected on the base and apex allow rotation and scaling to align each image on a normalized reference frame. B—Each rotated and scaled specimen image was assigned a weight, according to Equation 2. Weighted images were summed together. C—Weighted sums ix generated probabilistic distributions of where adipose was/was not on the ventricular surfaces...... 53

Figure 2.3: The modeled epicardial adipose distributions on the anterior right ventricle showed a low probability of coverage in an “L” shape on the right ventricular outflow tract. The apex of the left ventricle also elicited low adipose coverage...... 57

Figure 2.4: Diffuse regions of exposed myocardium were present on the modeled left ventricular margin. Epicardial adipose tissue occurred less on the basal posterior surface, and the central lateral wall of the left ventricle...... 57

Figure 3.1: (A) Shown here is a typical example of an axial slice of a high- resolution MRI scan of an isolated human heart. Regions with adipose tissue lying superficial to the ventricular myocardium are indicated with white arrows. (B) 50 anatomical reference points encircling both the left and right ventricles were selected. At each point, the relative adipose thickness was measured. Subsequently, the thicknesses at each defined anatomic location were averaged across of the entire sample of human hearts. TV = Tricuspid Valve, MV = Mitral Valve, PV = Pulmonary Valve, AIS = Anterior Interventricular Septum, PIS = Posterior Interventricular Septum ...... 70

Figure 3.2: High fidelity adipose reconstructions were generated from donated human hearts. (A) An image taken <24 hours post mortem shows the yellow adipose tissues on the anterior surface of human heart 212. (B) 3-Dimensional reconstructions of the adipose tissues were generated from high resolution MRI scans. Notably, the 3-dimensional model is reconstructed from an MRI scan after perfusion fixation. As such, the 3D model shows an analog of end diastole (filled heart), whereas the fresh imaging was taken of this heart unfilled. The reconstructions mirrored the fresh adipose distributions; again, differences in appearance sin these images result from the differences in volume load. (C) Adipose thickness maps demonstrate regions of high and low adipose thickness: here, regions are colored in a gradient from green to red representing adipose thickness of 1mm to >10mm, respectively. Gray regions showed no overlying epicardial adipose. This organ was donated from a 66 year old female, with no salient cardiac history. More information on this donated specimen may be found on the Atlas of Human Cardiac Anatomy free access website (www.vhlab.umn.edu/atlas/histories/histories.html Heart 212)...... 72

Figure 3.3: 3-Dimensional adipose reconstructions of the anterior right ventricles from 10 donated human hearts are shown. Here, the thickness of the superficial adipose layer is color-mapped, with thicker to thinner adipose regions indicated by a relative red-to-green color scale (depicted right, in mm); regions devoid of adipose are colored grey. Adipose deposits were thickest at the anterior interventricular borders and at the basal atrial-ventricular grooves. A majority of x the anterior right ventricles were covered with at least a thin layer of adipose. The thinnest regions of adipose on the anterior right ventricles followed an “L- Shaped” pattern; typically extending from the right ventricular apex, to the body of the right ventricle, through the right ventricular outflow tract, terminating at the base of the pulmonary artery. Specimens from left to right: (Top) Heart 099, Heart 097, Heart 131, Heart 141, Heart 198, (Bottom) Heart 165, Heart 140, Heart 084, Heart 251, Heart 229. Patient histories are available at the University of Minnesota’s Atlas of Human Cardiac Anatomy free access website (vhlab.umn.edu/atlas/histories/histories.shtml)...... 75

Figure 3.4: 3-Dimensional adipose reconstructions of the diaphragmatic inferior surfaces from 10 donated human hearts are shown. Here, the thicknesses of the superficial adipose layers are again color-mapped, with thicker to thinner adipose regions indicated by a relative red-to-green color scale (depicted right, in mm); regions devoid of adipose were colored grey. Each specimen elicited two elliptical regions flanking either side of the posterior interventricular septum which showed either thin adipose, or exposed myocardium. These two elliptical regions were the most consistent exposed/thin adipose regions from heart-to-heart, with the ellipses on the right ventricles generally being slightly larger than that of the left. Specimens from left to right: (Top) Heart 099, Heart 097, Heart 131, Heart 141, Heart 198, (Bottom) Heart 165, Heart 140, Heart 084, Heart 251, Heart 229. Patient histories are available at the University of Minnesota’s Atlas of Human Cardiac Anatomy free access website (vhlab.umn.edu/atlas/histories/histories.shtml)...... 75

Figure 3.5: 3-Dimensional adipose reconstructions of the left ventricular margins from 10 donated human hearts are shown. Here, the thicknesses of the superficial adipose layers were again color-mapped, with thicker to thinner adipose regions indicated by a red-to-green color scale (depicted right, in mm); regions devoid of adipose are colored grey. Adipose is thinner and scarcer on the left ventricular surface. Typically, the thinnest regions of adipose coverage were located along the posterior lateral wall of the left ventricle. Adipose distributions were thinner toward the base of the left ventricle (below the atrial- ventricular groove), and thicker toward the apex. Specimens from left to right: (Top) Heart 099, Heart 097, Heart 131, Heart 141, Heart 198, (Bottom) Heart 165, Heart 140, Heart 084, Heart 251, Heart 229. Patient histories are available at the University of Minnesota’s Atlas of Human Cardiac Anatomy free access website (vhlab.umn.edu/atlas/histories/histories.shtml)...... 76

Figure 3.6: Mean adipose thickness determined for the right ventricle across the evaluated 80 human hearts. Values re expressed in millimeters of adipose thickness; the color of each cell in the thickness map (left) corresponds to the mean thickness displayed, with red indicating thicker adipose coverage and green indicating thinner adipose coverage. Regions of interest on the right ventricle may be mapped between the anatomic reconstruction (right Heart 229) xi and the thickness map (left). Circled in green, was an elliptical zone on the diaphragmatic inferior surface of the right ventricle, which commonly showed the lowest average adipose thickness. Circled in orange, is a region on the right ventricular margin which commonly had the thickest adipose deposit on a given the right ventricle. Circled in black, is a zone on the right ventricular outflow tract; this region typically had the thinnest adipose layer on the anterior portion of the right ventricle. RVE = Right ventricular ellipse, RVM = Right ventricular margin, RVOT = Right ventricular outflow tract...... 77

Figure 3.7: Mean adipose thickness determined for the left ventricle across the evaluated 80 human hearts. Values are expressed in millimeters of adipose thickness; the color of each cell in the thickness map (left) corresponds to the mean thickness displayed, with red indicating thicker adipose coverage and green indicating thinner adipose coverage. Regions of interest on the left ventricle were mapped between the anatomic reconstruction (right, Heart 131) and the thickness map (left). Circled in orange, the anterior interventricular septum elicited thick adipose coursing along the left anterior descending coronary artery. The posterior interventricular septum, circled in red, had a correspondingly high adipose thickness. The basal posterior region of the left ventricular free wall, circled in black, displayed a lower average thickness than the overall left ventricular surface. The thinnest average adipose coverage on the left ventricle was found on the diaphragmatic surface of the posterior left ventricle, here, circled in green. AIS = Anterior interventricular septum, LVFW = Left ventricular free wall, LVE = Left ventricular ellipse, PIS = Posterior interventricular septum...... 78

Figure 3.8: Mean adipose thickness varied by anatomic location. On this population of 80 human hearts, the regions with the thinnest adipose depositions were two elliptical regions on either side of the posterior interventricular septum. The regions with the thickest adipose coverage were typically located on the right ventricular margin and the anterior and posterior interventricular septa. Anatomic subsections, as quantified here, were defined in Figure 3.6 and Figure 3.7. Error bars represent the standard error of the mean. RV Ellipse = 1.14±0.82mm, LV Ellipse = 1.27±0.64mm, LVFW = 1.40±0.46mm, LV Overall = 2.02±0.28mm, RVOT = 3.24±0.64mm, RV Overall = 4.73±0.28mm, PIS = 6.78±0.50mm, RV Margin = 7.44±0.57mm, AIS = 9.01±0.50mm. RV = Right Ventricle, LV = Left ventricle, LVFW = Left ventricular free wall, RVOT = Right ventricular outflow tract, PIS = Posterior interventricular septum, AIS = anterior interventricular septum...... 79

Figure 4.1: Medtronic 3830 leads were placed in distinct zones along the atrial septum. Implant regions were defined by their proximity to the coronary sinus and antero-septal commissure. 1) proximal to the coronary sinus, 2) midway between the antero-septal commissure and coronary sinus, 3) proximal to the

xii antero-septal commissure, 4) below the tricuspid annulus near the antero-septal commissure...... 94

Figure 4.2: Pacing capture strength duration curves vary minimally as a function of implant location along the bundle of His. Capture threshold is likely a function of proximity and orientation relative to the bundle of His—variables not explicitly tested in this experimental paradigm. HB = His Bundle ...... 95

Figure 4.3: The ratio of QRS width between paced and native sinus beats (∆QRSpaced/∆QRSnative) increases when moving distal along the bundle of His, suggesting that proximal bundle pacing results in a more physiologic response, **p<0.005. HB = His Bundle...... 96

Figure 4.4 (Top) Evoked QRS complexes in Lead II for implants in locations 2 through 4, as compared to the native QRS complex. All signals resulted from VOO pacing. In location 1, no HBP capture was achieved. In location 2, selective HBP capture resulted in an ~40ms isoelectric period prior to ventricular depolarization, yielding identical QRS morphology to the native beat. As leads are implanted progressively more distal along the HB (locations 3 and 4), evoked QRS becomes wider, and non-selective HBP capture is achieved. In location 2, the timing between the pacing spike and the end of the QRS was 98ms. In location 4, the timing between pacing spike and the end of the QRS was 106ms. This suggests primary right bundle capture, followed by retrograde left fascicular activation. (Bottom) Histologic sections for the leads implanted in locations 2, 3, and 4; yellow arrows indicate lead implant sites. In location 2, the lead was implanted directly onto the proximal HB. In location 3, the lead was implanted onto the penetrating HB, just prior to bifurcation. In location 4, the lead was implanted past the bifurcation of the HB, nearest to the RBB, confirming the hypothesized RBB pacing capture to retrograde LBB activation pattern as described above. HB—His bundle; LBB—Left bundle branch; RBB—Right bundle branch ...... 97

Figure 4.5: Native His bundle electrogram timing with respect to QRS complex varies as a function of length along the bundle of His. Moving distally from the coronary sinus toward the antero-septal commissure, timing between His spike and the V-wave decreases. Notably, in region 1, native conduction system tissues may or may not be present, thus the His potential may or may not be present. The pictured electrical signals are an artistic rendering of lead electrograms, as opposed to actual recordings. Here, in each of three implant locations on the atrial side of the tricuspid valve (locations 1, 2, and 3), a significant atrial EGM component can be seen. CS=coronary sinus, IVC = inferior vena cava ostium, RA=right atrium, H= His potential electrogram, V = Ventricular electrogram...... 98

xiii

Figure 4.6: A—The time interval between the His potential and the takeoff of the QRS complex in the native electrogram decreases moving distally along the bundle of His. In region 1, His to QRS interval was the largest, averaging 49ms. In region 4, below the tricuspid annulus, His to QRS timing averaged 35ms. B— The time interval between pacing spike and QRS takeoff decreases moving distally along the bundle of His. Here, this time interval is defined as the time between pacing spike and the end of the delta wave, so to isolate Purkinje conduction. As such, leads implanted in region 4, below the tricuspid annulus, show a shorter conduction time between pacing spike and ventricular contraction. Conversely, leads implanted in regions 1,2, and 3 above the tricuspid annulus show longer intervals from pacing impulse to QRS initiation. . 99

Figure 4.7: Lead locations and corresponding histology. Image A: Lead locations 1 through 4 within the right atrium in Dog 4. Images B-E: Histologic sections at low magnification (1x) of each lead implant site (indicated by arrows). Masson’s trichrome stain. Bar = 2mm. Higher magnification (40x) images below (Images F- I) correspond to the yellow boxed region in the above Histologic sections (Image B-E). Image B & F: No major conduction structure identified in proximity to the implant site of Lead 1. Image C & G: The atrioventricular node (AVN) is nearest to the implant site of Lead 2. Image D & H: The His bundle (HB) is nearest to the implant site of Lead 3. Image E & I: The right bundle branch (RBB) at the His bundle bifurcation is nearest to the implant site of Lead 4. Ao=Aorta, AVN=atrioventricular node, CS=coronary sinus, IVC=inferior vena cava, HB=His bundle, LA=left atrium, LBB=left bundle branch, LV=left ventricle, RA=right atrium, RAA=right atrial appendage, RV=right ventricle, TV=tricuspid valve .... 100

Figure 4.8: Histology sections from implant region 2 (A) and implant region 4 (B). Yellow arrows depict the implant site of the lead being analyzed. Within region 2, the main body of the proximal bundle of His, near its junction with the compact atrioventricular node, is accessible for pacing therapy. Region 4, just below the tricuspid annulus sits at the branch point of the bundle of His into the right and left bundle branches. HB = bundle of His, LBB = Left Bundle Branch, RBB = Right bundle branch...... 103

Figure 4.9: A--If a lead is implanted in region 2, it may activate the proximal bundle of His, leading to more physiologic conduction. B--A lead implanted in region 4 may activate only a portion of the His-Purkinje system, leading to less physiologic pacing. C&D—Notably, the optimal location to pace along the bundle of His may vary depending on an individual patient anatomy. As depicted here, the presence of fascicular block in the left bundle branch may remove the conferred benefit of proximal bundle pacing...... 106

Bipolar 3830 recording for native sinus beats. Atrial and Ventricular EGMs were taken from temporary pacing leads placed in the right atrial appendage and the right ventricular outflow tract. CS—Coronary Sinus; TV—Tricuspid Valve; MS— Membranous Septum ...... 116

Figure 5.2: A&B) Endoscopic footage of 3830 lead implants on the atrial and ventricular side of the tricuspid annulus. Both leads were placed along the lower border of the membranous septum. C&D) Native electrical signatures for each lead placement. Traces from top to bottom: ECG Leads I, II, III, AEGM, VEGM, 3830 bipolar EGM. E&F) Selective HBP capture was achieved for both lead placements. Traces from top to bottom: ECG Leads I,II,III, AEGM, VEGM, 3830 bipolar EGM...... 118

Figure 6.1: The right atrial appendage of the human heart...... 125

Figure 6.2: (right) A business card is read through the body of the right atrial appendage. Thicker pectinate muscle bands span the clear visceral pericardial layer...... 126

Figure 6.3: High resolution MRI slice from a human heart at the University of Minnesota's Visible Heart Lab ...... 126

Figure 6.4: Highlighting the vestibule (non-pectinated) region of the right atrium. (Right) Inferiorly, the right atrial appendage borders the annulus of the tricuspid valve. Pectinate muscles extend radially from the annular ring. (Left) On the septal side of the crista terminalis, there are no pectinate bands to affix to...... 127

Figure 6.5: Highlighting the pectinated region. So to not artificially increase the calculated pectinate surface area, the epicardial surface of the pectinated appendage was highlighted, thus excluding the additional surface area from the mounded pectinate muscle...... 128

Figure 6.6: Measuring pectinate muscle within the right atrial appendage. Pectinate thickness was evaluated in both a long and short axis. Pectinate spacing was defined as the distance to the closest major pectinate muscles on either side of the muscle being analyzed...... 130

Figure 6.7 Pectinated surface area within the human heart increases with age and overall patient height and weight. However, the proportion of pectinated to overall surface area within the right atrium is relatively invariant with these patient factors...... 133

Figure 6.8 Pectinate size and the spacing between the closest major pectinate muscles varies as a function of location in the right atrial appendage. Moving anteriorly from the lateral wall to the right atrial appendage tip, pectinate become more thinner and more densely packed...... 134 xv

Figure 7.1 A: A custom-made chamber secured the RAA during penetration tests. B: Penetration of tissue is a combination of applied force, and the displacement (tenting) applied to the tissue. The methodology in this experiment is based on protocols designed by Eggen et al...... 143

Figure 7.2: Example perforation test for the right atrial appendage. The first peak in the force vs displacement trace represents the force and distance required for pectinate muscle perforation. The second, larger peak shows the force and distance needed for perforation through the epicardial surface of the right atrial appendage...... 144

Figure 7.3: The relationship between penetrator surface area and force at tissue penetration. N On average, epicardial (thin wall) tissues take a greater amount of force to penetrate than pectinate muscle...... 145

Figure 7.4 Cumulative probability distributions of perforating through a given human heart tissue layer as a function of displacement and pressure applied. At a pressure of 5N/mm2, 90% of all pectinate muscles were penetrated, while only 50% of epicardial layers were perforated. Similarly, at 5mm of penetrator displacement, 60% of all pectinate layers were penetrated, while no perforations were seen...... 147

Figure 8.1 A typical active fixation lead within the right atrial appendage. Atlas of Human Cardiac Anatomy ...... 154

Figure 8.2: The experimental setup for atrial stiffness testing. A suture was attached to the body of the right atrial appendage; the suture passed through the tricuspid valve, and through a hemostasis port in the right ventricular apex. Subsequently, the suture was threaded over a pulley and attached to a linear force transducer, which measured tensile force as a function of applied displacement...... 156

Figure 8.3 Example force vs displacement curve from stiffness testing. Force linearly increases (slope=stiffness) as a function of displacement. Notably, in each individual test, cyclic variation can be seen due to the beating of the right atrial tissue. Here, the peaks in the cycle represent atrial diastole, while the troughs represent atrial systole...... 157

Figure 8.4: Right atrial appendage compliance decreased approximately linearly with right atrial pressure. Accordingly, at higher atrial pressures—as could be predicted in patients with pulmonary hypertension, significantly more force is required to displace a given distance...... 158

Figure 8.5: Right atrial force of inversion varied with right atrial pressure. As mean RAP increases, inversion force increased approximately linearly...... 159

xvi

Figure 9.1 : (A) Encapsulation is visualized at the distal end of the MicraTM, (B) The MicraTM is retrieved, (C) Post-extraction, the fibrous capsule remains...... 167

Figure 10.1: There is no clinically-available retrieval tool for the MicraTM transcatheter pacing system (TPS). As such, off-the-shelf catheters have been utilized for retrieval. Here, a steerable sheath is used in combination with a foreign body retrieval snare to capture the retrieval feature of the MicraTM device. The device is then extracted, pulling the fixation tines through the tricuspid valve...... 176

Figure 10.2 Benchtop testing paradigms to test biomechanical properties of the chordae tendineae. (A) Leadless pacemaker fixation tines were engaged with chordae tendineae of the tricuspid valve; fixation tines were extracted via a 120mm/min uniaxial pull. (B) Biomechanical properties of isolated chordae tendineae were derived. Isolated chordae underwent a 120mm/min uniaxial tensile test until tissue failure...... 178

Figure 10.3 Chordal failure forces greatly exceed the forces required to extract engaged leadless pacemaker tines. Force values are normalized to the mean force of chordae tendineae failure during 3-point tensile testing. Notably, during tine extraction testing, no chordal failure was seen. (N=6 human hearts, n=60 tissue failure tests, n=50 single tine extrication tests, n=50 two tine extrication tests) ...... 181

Figure 10.4 Chordal failure forces vary across papillary complexes. Failure occurs at a lower force when testing the septal papillary complex. Notably, this represents failure of the entire chordal complex (i.e. including leaflet and papillary attachment points). Forces are normalized to the mean force of chordae tendineae failure seen during 3-point tensile testing. (N=6 human hearts, n=60 tissue failures) ...... 182

Figure 10.5: During three-point failure testing, failure rarely occurred in the body of the chordae tendineae. Most commonly, the chordal complex failed at the junction between the chordae and the tricuspid valve leaflet and/or the junction of the chordae tendineae with the papillary muscle. (N=6 Hearts, n=60 tissue failures) ...... 183

Figure 10.6: The stress and strain required to rupture the body of the chordae tendineae varies by papillary attachment location. Overall, septal complex chordae required significantly more stress and strain to rupture. (N=4 human hearts, n=37 uniaxial chordal failures) ...... 184

xvii

Section I. An Introduction to Cardiac Pacemakers: Where have we been? Where are we now? What’s next on the Horizon?

The first section of my thesis lays the foundation for the motivations of my research. The cardiac pacing field as we know it today is the result of several decades of discovery. In order to continue pushing pacing therapy forward, it is important to understand the groundwork that has been laid and where the next wave of innovation might lead.

Chapter 1: The Cardiac Pacemaker: A Crossroads of Engineering and Medicine

Alexander R. Mattson BS1,2, Michael D. Eggen PhD2,3, Paul A. Iaizzo PhD1,3

1Department of Biomedical Engineering, University of Minnesota, Minneapolis,

MN USA

2Medtronic, Mounds View, MN USA

3Department of Surgery, University of Minnesota, Minneapolis, MN USA

1.1 Preface Over the past 70 years, the cardiac pacemaker has blossomed into a life- saving technology, with hundreds of thousands of new implants each year.1

Despite a rich history, the cardiac pacing industry remains on the cutting edge of innovation, pushing boundaries to reduce complications of current therapy and extend pacing therapy to new patient populations.

With innovation come significant challenges. Each new pacing technology requires an intricate understanding of the anatomy, physiology, and biomechanics of the tissues in which it will be implanted. This thesis tackles some of the central questions surrounding emergent pacing technologies.

In order to gain an in-depth understanding of pacemaker technology and where future innovation may lead, it is important to learn valuable lessons from past innovators. This chapter dives into the rich history surrounding the field of cardiac pacing, providing the foundation for the remainder of the thesis. It describes the current technology standards, discusses different pacing paradigms, and offers a glimpse into the bright future of cardiac pacemaker therapy.

This chapter was written and edited by Michael Eggen, Paul Iaizzo, and me. The chapter will be published as chapter 6 of the textbook “Engineering in

Medicine: Advances and Opportunities.”

1.2 Looking back: A brief history of cardiac pacing

Lidwill, Hyman, and Zoll: Forefathers of Cardiac Pacing At its inception, electrical stimulation of the heart was thought only to serve the purpose of resuscitating patients who had experienced a recent sudden cardiac arrest. Interestingly, prior to the late 1920s, cardiac arrest was a relatively common, and frequently fatal occurrence in the operating room.

Resuscitation attempts centered around delivering epinephrine to prime the heart, coupled with mechanical excitation of the ventricular myocardium. It seemed that after arrest, the threshold for electrical conductivity through the heart was lowered; the prick of a needle was occasionally sufficient to send a sweeping electrical pulse through the myocardium, restoring cardiac rhythm.

Mark Lidwill and Albert Hyman, early pioneers of pacing, paired these resuscitation efforts with their basic understanding of the ECG and surrounding electrophysiology.2,3 Their efforts centered on electrically stimulating the atrial half of the heart, hoping that upon cessation of a brief external stimulation, the heart would resume a native sinus rhythm, and begin beating again. The first documented evidence of using electrical stimulation to stimulate the human heart was provided by Lidwill in 1929.2 Shortly after in 1932, Hyman published his electrical resuscitation efforts.3

Figure 1.1 Albert Hyman's original device. This "pacemaker" was designed to provide electrical shocks at the turn of a hand crank, revitalizing a stopped heart.

Unfortunately, Lidwill and Hyman’s inventions did not immediately flourish.

The medical community at large, was skeptical toward their work: many physicians could not palate the treatment of cardiac arrest, due to low success rates of resuscitation and the inherent invasiveness of placing a plunged needle electrode into the heart of a patient near death. Unbeknownst to the inventors, cardiac arrest predominantly results from arrhythmia known as ventricular fibrillation, a condition which pacing cannot cure—thus, they ought to have invented the defibrillator! The true promise of cardiac electrical stimulation was not unearthed until the 1950s.

1.3 Lillehei and Zoll: World War II Innovators Prior to 1950, the heart was thought to fundamentally drive human emotion; to some, it was the center of the soul itself. There were no methods for conducting external heart surgery and no way to sufficiently oxygenate the brain 5

during surgery. As such, all efforts to surgically repair the heart were made while it remained beating. Any stoppage of blood flow during these procedures could result in damage to the brain; hence, only simple surgeries could be performed on the beating heart.

This thinking began to change, when World War II inspired a surge in surgical advances; daily life vs death operations were performed by field surgical teams, spurring rapid innovation. Surgical teams returning from the war were armed with new surgical techniques and an inspired confidence to take on any challenge they were presented with. One of these young war-experienced surgeons, C. Walton Lillehei, returned to the University of Minnesota in 1950 to complete his surgical residency after leading an Army surgical field unit in both

North Africa and Italy. Lillehei, oft described as a maverick, pushed the next level of care for his patients by rebelling against what convention might propose for him. Notably, Lillehei played a fundamental role in the development of deep hypothermia, cross-circulation, bubble oxygenation, and revolutionary surgical techniques to repair congenital heart defects. Unfortunately, his pioneering surgical work was not without complication. Early cardiac surgery pioneers were not in-tune with the challenges of working with the human cardiac conduction system. In 10-20% of Lillehei’s early cases, children would develop complete heart block, a fatal consequence at the time. To solve this dilemma, Lillehei looked across the country, and drew inspiration from the work of Dr. Paul Zoll.

While Lillehei was pioneering cardiac surgery in Minnesota, Dr. Paul M

Zoll, a research cardiologist at Harvard Medical School, was laying the groundwork for clinical pacing therapy. Zoll’s pacemaker concept strapped two pacing electrodes on opposite sides of the thoracic cavity, such that the generated electric field would pass through both ventricles of the heart.4 The group could feasibly control the heart rate, effectively eliminating the arrest brought on by heart block. By 1952, Zoll had begun using his device on human patients. In one 65-year old patient, Zoll’s device was able to maintain a stable heartbeat for more than 50 hours at a time.4 Unlike the preceding Lidwill and

Hyman devices, Zoll’s approach was non-invasive, using stimulation electrodes placed on the skin, as opposed to plunge electrodes placed within the myocardium. Zoll’s method also stimulated the ventricular half of the heart, and as such could treat those patients affected by AV nodal block. His pioneering research spurred innovation across the medical community, leading many to look at electrical stimulation as a viable means for treating cardiac ailments.

Lillehei knew of Zoll’s successes across the country but was skeptical that the Zoll pacemaker could fit his needs. Zoll’s device was designed to pace non- invasively; thus, it required large voltages to capture ventricular myocardium through the thorax. Patients often complained of the pain of repeated high- voltage electrical stimulation, often capturing thoracic skeletal muscle in addition to cardiac tissues. Lillehei needed an improved concept for his patients.

1.4 Earl Bakken and the Birth of a Medical Device Industry Lillehei’s surgical pacing approach employed a multistrand, braided stainless steel wire implanted directly into the ventricular myocardium, with the other end brought through the surgical incision and attached to external stimulation. This method of treatment, suggested by Dr. John A. Johnson, a professor of Physiology at the University of Minnesota, required only small magnitude electrical stimuli to effectively control heart rate.5 The initial pulse generators for Lillehei’s pacemakers were plugged into an external wall socket, just like those that Zoll used, limiting patient mobility to the length of an extension cord. With a patient population dependent on the functioning device, Lillehei ran the risk of power outages endangering his patient’s lives. On October 31, 1957, that is exactly what happened. A sudden city-wide blackout tragically ended the life of an infant patient. Lillehei, again, needed an innovative solution.

At the time, the University Hospital subcontracted with a local electric equipment repair company, Medtronic Inc., to perform maintenance tasks on operating room equipment. At the time, Medtronic was a two-person company—

Earl Bakken and his brother-in-law Palmer Hermundslie.5 Bakken was a trained electrical engineer who received his degree from the University of Minnesota; importantly, he was present during the cardiac surgical procedures in the

University’s operating rooms. Lillehei approached Bakken to see if he was up to the task of designing an improved pacing system.

A B

Figure 1.4: Earl Bakken's first transistorized pacemaker was developed over a course of weeks (A), animal tested over a course of hours (B), before first-in-human implantation.

Bakken began this work in 1957, using a circuit modified (in his words

“plagiarized”) from a circuit diagram for a transistorized metronome described in a Popular Electronics magazine.6 In 1958, Bakken brought the pacemaker to the

University’s animal laboratories to test the device on a dog (Figure 4 ). The device functioned as intended, and so, Bakken returned home for the evening.

On April 14, 1958, the “battery-powered, wearable pacemaker” was first used clinically, even though this was somewhat unplanned. In the words of Bakken, 10

“The next day I returned to the hospital to work on another project when I happened to walk past a recovery room and spotted one of Lillehei’s patients. I must have done a double take when I glanced through the door. The little girl was wearing the prototype I had delivered only the day before!”6 Bakken’s transistor pulse generator had made a miraculous “overnight” transition from pre- clinical animal testing to clinical use.7 Bakken commercialized his device, recessing the dials, and streamlining the design. Soon he was receiving pacemaker orders from all over the country, and Medtronic became a dedicated pace-making company, which would eventually blossom into one of the largest medical device companies in the world.

1.5 Innovation Fuels Innovation The kick start to cardiac pacing was not without obstacles. By wiring an external device through a surgical incision, surgeons created a source for infection, which needed constant monitoring. Several groups considered fully- implanted devices to alleviate these problems.

Figure 1.5: A patient wearing a Medtronic model 5800 pacemaker. The first commercialized pacemakers were worn around the neck. There were common clinical complications with the lead entry sites.

In Europe, Ake Senning and Rune Elmqvist, a Swedish surgeon-engineer pairing, made the first attempt at an implantable pulse generator. The initial device implant lasted only three hours, and a second attempt failed after a few weeks. Interestingly, the patient who was given this preliminary implantable device lived for another 44 years, receiving over 25 new pacemakers along the way.8

Across the Atlantic, unaware of Senning and Elmqvist’s efforts, another physician-engineer pairing, Dr. William M Chardack and Wilson Greatbatch, took a different approach to creating an implantable device. Greatbatch created a device with a mercury-zinc battery cells, and shrunk the device circuitry into a package slightly smaller than a hockey puck.9 Chardack and Greatbatch’s first prototypes suffered from a phenomenon seen in many of the early Medtronic 12

Model 5800 external pacing systems: threshold rise, and subsequently greater power consumption. While external pacemakers could simply switch out batteries, rising threshold was a death knell for implanted systems.

Fortunately for Chardack and Greatbatch, Dr. Samuel Hunter had been collaborating with Hunter Roth, a Medtronic Engineer; the duo had devised an innovative solution to early pacing threshold challenges. The Hunter-Roth electrode required roughly 70% less current to pace the myocardium than existing electrode technology.10 The bipolar electrode partially solved Chardack

and Greatbatch’s power consumption issues, allowing them to license their technology to Medtronic, who manufactured and sold 50 of the devices within the year. The device line remained the most widely-used pacemaker throughout the

1960s.

Figure 1.7: Pacing electrodes invented by Dr. Samuel Hunter and Medtronic engineer Hunter Roth. The stainless steel Hunter-Roth electrode helped to reduce pacing capture threshold.

Figure 1.8: Dr. Samuel Hunter (inset) and Medtronic engineer Norman Roth developed a bipolar electrode that represented a major advance in pacing technology. First implanted in 1959, the

Hunter-Roth lead helped contribute 7 years of life to a Stokes-Adams disease patient, Warren Mauston in 1960.

Around the same time frame, Dr. Seymour Furman was tasked with setting-up an open-heart surgical program at New York’s Montefiore Hospital; this included, in part, learning the technique of cardiac stimulation pioneered by the Minnesota surgical teams. Concurrently, Furman was learning cardiac catheterization. Furman connected the dots and began efforts to pace the ventricle via a subclavian catheterization. Furman’s second patient, an elderly gentleman of 76 years was a pivotal success story. By threading a catheter through the left subclavian vein, Furman paced the man’s right ventricle at a threshold of 1.5-3.0V for 96 days.11

Furman’s technique resolved many issues that plagued Lillehei’s myocardial pacing method. Namely, it did not require a thoracotomy, making the therapy available to sick patients who otherwise could not tolerate the surgery.

Additionally, pacing capture was achievable at the low voltage, and did not appear to suffer from the same rising phenomenon seen in myocardial pacing. In

1965, Medtronic released a pacemaker intended to be used with a transvenous lead, and transvenous pacing rapidly became the standard of care. Still today, the vast majority of pacemaker implants employ a transvenous lead.

The early days of pacing, systems were marked with rapid improvements on all fronts. In the early days of pacing, mercury-oxide zinc batteries resulted in limited device longevity. At a low pacing capture threshold, these batteries lasted

around two years. Longer-lasting, reliable batteries were needed for successful implantable technologies. Fortunately, in the 1960s and early 1970s, the medical device industry latched on to research on lithium ion batteries that the military had been researching for years. Wilson Greatbatch, already famous for his work in implantables, brought lithium batteries into the medical space in 1971, and only a few years later, the first lithium ion pacemaker hit the medical market.

Due to their vastly increased longevity, lithium ion batteries rapidly ascended into prominence, and by 1978, their uses comprised a majority of pacemakers on the market.

In the subsequent decades, pacing therapy went through a series of improvements, while maintaining similar core technologies. Transistor-driven pacemakers were replaced by integrated circuits, which eventually were replaced by microprocessors.

1.6 Modern Implantable Pacing Systems Today’s Pacing Systems Today, implantable pacing systems consist of multiple components working as a unit to deliver effective therapy. Generally, the chronically implanted components of a modern pacing system include the implantable pulse generator

(IPG, or pacemaker) as well as the pacing lead or leads. The IPG can is most commonly implanted subcutaneously near the left pectoral region, but may be shifted to the right pectoral region if the left subclavian vasculature is unsuitable, if other devices are present, or if there is an overarching patient or physician preference (e.g. left handed trap shooters may wish to have a right pectoral 16

implant to avoid interaction on their shooting shoulder). IPGs may also be placed submuscularly, so to prevent or minimize subcutaneous pocket erosion and/or for cosmetic considerations. In pediatric patients they are often implanted abdominally.

Today, pacemaker rates and modes of function, can be programmed via telemetry. An external programmer telemeters information to and from the programmable IPG. As such, the cardiac electrophysiologist may change pacing parameters and download electrical information captured by the device without direct access to the IPG.

It should be again noted, that early pacing systems utilized plunge electrodes placed through the epicardial wall of the heart. In these systems, it was common to implant the pulse generator in the abdomen, due to the large size of the can, and easier access to the epicardial surface. This technique is still used in certain clinical situations (e.g., neonates, or individuals with compromised vascular access). However, a transvenous lead approach, as pioneered by Furman, is far more common.

Transvenous pacemaker leads are generally placed via one of two methods: the surgical cephalic cutdown, or a subclavian puncture. In a surgical cutdown, the vessel is exposed via careful dissection, and a small incision on the vessel wall allows direct insertion of the lead. In a subclavian puncture, a needle passes into the lumen of the vessel, allowing guidewire introduction.

Subsequently, a percutaneous lead-introducing catheter is placed over the wired- 17

rail and into the venous lumen. The guidewire is removed and the lead is then inserted through the catheter and advanced into the right side of the heart, allowing fixation in the right atrium, right ventricle, and/or through the coronary sinus to a cardiac vein on the left side of the heart: all depending on the type of system chosen, single, dual or biventricular systems. An anchoring sleeve secures the lead at the entry site, preventing external forces from dislodging or displacing the implanted lead(s). The lead terminals are connected to the respective inserts of the IPG. Generally, the IPG will be placed in a subcutaneous pocket near the left or right pectoral. Suturing closes the incision, completing the implantation.

1.7 Implantable Pulse Generators (IPGs) In a modern pacing system, an implantable pulse generator is controlled by a small microprocessor computer housed within a hermetically-sealed pacemaker housing, typically referred to as the “can.” The circuitry is generally powered by a lithium ion battery, which routinely takes up more than half of the can volume. In pacing systems, device longevity varies by utilization (i.e. pacing capture threshold, circuit impedance, and their associated current drains), but commonly exceeds 8-10 years. The internal circuitry is wired to an external connector block, which allows the transvenous pacing leads to interface with the internal device circuitry.

1.8 Sensing Algorithms Modern pacing systems have evolved significantly from the basic functionality seen in early pacemaker prototypes. Within the IPG, microprocessors combine to both sense electrical activities and deliver timed pulses, programmable based on the desired therapy.

Delivering appropriate therapeutic intervention, relies heavily on accurate detection and interpretation of intrinsic electrical signals within the heart. The pacing system will measure a voltage between the bipolar electrode pair on the 19

lead (bipolar lead) or between the lead tip cathode and the IPG (unipolar lead); this signal is known as the cardiac electrogram (EGM). Within the IPG, the hearts electrical activity is computationally interpreted by sensing algorithms.

Most rhythm management decisions are based on the patient’s detected heart rate. The modern IPG continuously measures the time from one sensed event to the next and compares the interval to the rates and intervals programmed by the clinician. For example, if two atrial events occur with a separation of 1,500 milliseconds (1.5 seconds), the heart rate is 40 beats per minute (HR=60/measured beat-to-beat interval; 60/1.5=40 beats per minute).

1.9 Pacemaker Leads

Cardiac pacing leads are the traditional electrical conduit between the implantable pulse generator and the heart. Leads allow transmission of electrical energy to the heart to artificially stimulate contraction, and also relay sensed electrical signals back to the implanted pulse generator for diagnostics and cardiac monitoring. Pacemaker leads commonly traverse through diverse, dynamic vascular and cardiac anatomies. As such, leads must be able to withstand the foreign body response of the human body, anchor securely within myocardium while maintaining the flexibility necessary to navigate through venous and cardiac anatomies, and undergo approximately 400 million cardiac cyclic deformations over a 10-year period without fracture or damage.

Currently, leads are most commonly placed endocardially within the cardiac chambers, but may be placed on the heart’s external surface,

“epicardially,” depending on the patient’s indication, physician preference, and/or anatomic considerations. Endocardial leads generally originate from the implantable pulse generator in the subcutaneous space, enter the subclavian venous vasculature, travel through the superior venous return, and anchor their distal tips within the myocardium. Again, today, the vast majority of pacing systems utilize endocardial leads, taking a subclavian venous approach, as

originally described by Furman. In contrast, epicardial leads are attached directly to the surface of the heart: these leads are most commonly used when transvenous implantation is not feasible; i.e. in pediatric patients or adults with compromised and/or tortuous venous vasculature.

Modern leads are all constructed with biocompatible polymers and biostable metals. Yet, lead body design is variable based on the number of circuits required, its ultimate size (e.g., diameter or French size and overall length), handling considerations, and manufacturer preferences.

At the distal tip of the pacing lead, an electrode interfaces with cardiac tissue, must provide stable, chronic electrical sensing and stimulation. Electrode configurations also vary across commercially available leads, designed in both unipolar and bipolar configurations. Unipolar pacing circuits utilize a single cathodal electrode at the tissue interface, with the implanted pulse generator serving as the anode. Bipolar pacing systems use an electrode at the tissue interface as a cathode but have a pacing anode positioned on an adjacent section of the distal lead body. Pacing leads commonly use a cylindrical electrode placed along the lead body (ring electrode) as the anode.

To facilitate stable electrical interface at the cathode, pacemaker leads will often utilize an active or passive fixation mechanism at the lead’s distal tip.

Passive fixation mechanisms do not require and active deployment by a clinician: passive mechanisms commonly include polymeric tines and/or shaped segments along the length of the lead. Conversely, active fixation mechanisms require

some active deployment from the clinical implanter; e.g. deployment of helices, hooks, or barbs. Additionally, some epicardial leads require sutures to maintain a stable position.

Figure 1.11: Pacing leads come equipped with either passive or active fixation. Shown here are a tined passive fixation and a helical active fixation (cork-screw) mechanism.

Typically, the distal portion of the lead is designed to interface with the cardiac tissue. Thus, in order to optimize chronic electrical performance, the distal lead tip must: 1) minimize inflammatory response, 2) provide high capacitances and impedances, and 3) act as a stable fixation mechanism throughout the life of the lead. Most modern cathodic electrodes elute anti- inflammatory agents/drugs (e.g., dexamethasone sodium phosphate) to suppress inflammatory responses; this helps to manage acute changes in the cardiac tissue at the implant site, stabilizing chronic pacing and sensing performance.

Many pacing electrodes are also given coatings to produce a large, highly capacitive surface area; this reduces battery drain, and results in small polarization following a pacing pulse, increasing sensing performance. However, the size of the pacing cathode has decreased over time, increasing cathode- tissue impedance and augmenting system efficiency by reducing the current drain of each pacing impulse.

1.10 Clinical Pacing Categories Bradycardia Pacing Pacemaker therapy is most commonly used for symptomatic bradycardia

(an abnormally slow heart rate). Typical causes of these bradyarrhythmia are:

1) sinus node dysfunction; 2) acquired permanent or temporary atrioventricular block; 3) chronic bifascicular or trifascicular block; 4) hypersensitive carotid sinus syndrome; 5) neurocardiogenic in origin; and/or 6) a side effect due to a drug therapy. Pacemaker therapies for symptomatic bradycardia comes in a variety of forms, including: ventricular pacing, atrial pacing, and dual-chamber pacing. The prescribed therapy is typically dependent on an individual patient’s pathology, age, and previous medical history. For example, a patient with AV nodal block might require ventricular pacing to appropriately time the ventricular contraction with atrial depolarization. Conversely, a patient with sinus node dysfunction, but an otherwise unaffected conduction system, might require only atrial pacing.

In the early 1990s, advances in lead and pulse generator technology led to dual-chamber pacing systems. New polyurethane leads made for easier implantation of multiple leads, and allowed right atrial (RA) and right ventricular 24

(RV) pacing to be accomplished in the same patient.12 Further, advanced tined lead designs dramatically reduced atrial lead dislodgements.13 Similarly, the aforementioned developments in integrated circuits and pacemaker microprocessors led to IPGs with sophisticated algorithms for novel treatments and cardiac monitoring. As such, IPGs may be programmed to deliver both atrial and ventricular pacing in patients with chronotropic incompetence that necessitates pacing in both atrial and ventricular chambers of the heart.

Cardiac Resynchronization Therapy Cardiac pacing has historically been used to treat bradycardia (delayed or absent activation of the entire ventricle). The conception of cardiac resynchronization therapy (CRT) came from two separate studies. A French clinical research team showed that multisite atrial pacing--one lead in the right atrial body, a second in the coronary sinus—aided in suppressing atrial tachyarrhythmias (AT) in patients with an intra-atrial conduction delay.14 This therapeutic innovation led to the development of a dedicated coronary sinus pacing lead: the model 2188 CS lead (Medtronic plc, Minneapolis, MN, USA). In

1994, a patient with severe congestive heart failure received a four-chamber pacing system utilizing standard right atrial and ventricular leads, the model 2188

CS lead, and an epicardial Medtronic 5071 lead on the left ventricular free wall.15

The patient’s clinical status improved dramatically, showing the feasibility of multi-chamber resynchronizing pacing.

These original research studies led to the clinical development and use of cardiac resynchronization therapy (CRT). At its core, CRT improves cardiac function by restoring coordinated (more physiologic) contraction across the left ventricle. In some cardiac pathologies—commonly left bundle branch block— segments of the left ventricle of the heart shows marked delay in activation. This results in inefficient contraction of the ventricles, leading to reduced stroke volumes and cardiac outputs. Broadly, this left-sided delay with respect to the right is termed dyssynchrony. Cardiac resynchronization involves pacing a single locus (or multiple loci) within the left ventricle to improve ventricular electrical

(and consequently mechanical) coordination. Generally, CRT therapy involves placing a pacing lead through the left coronary venous vasculature, resulting in epicardial pacing of the left ventricle. This lead will provide stimulus nearly simultaneously with a secondary lead placed in the right ventricle. By resynchronizing conduction, immediate acute improvements in a patient’s left ventricular hemodynamic function can be seen—i.e. increased ventricular contractilities, stroke volumes, and ejection fractions.16,17 In selected patient populations, CRT has been shown to reverse pathological chamber remodeling, improve ventricular function indices, reduce traditional heart failure biomarkers, and lowering mortalities.18–20 Yet, successful CRT is considered to be dependent on careful patient selection. Further, it is critical that the resynchronization lead be placed in the optimal position and the device be programmed appropriately to normalize left ventricular conduction.

His Bundle Pacing

Traditional ventricular bradycardia pacing utilizes leads placed within the endocardial body of the right ventricle. While traditional pacing therapies are effective, right ventricular (apical) pacing inherently causes interventricular dyssynchrony and adverse hemodynamics, as the pacing-induced ventricular depolarization does not utilize the native cardiac conductive system. Some have proposed an alternative therapy: His bundle pacing. The bundle of His is the beginning of the ventricular conduction system. It originates from the distal 27

portion of AV node, and travels through the central fibrous body before bifurcating into the left and right bundle branches. Thus, by pacing the bundle of

His, the pacing-induced depolarization waves will utilize the native cardiac conduction system to travel through the myocardium, producing theoretically ideal ventricular synchrony. Permanent His bundle pacing was first described in

2000, in a subset of patients with chronic atrial fibrillation and dilated cardiomyopathy.21 Across 18 patients, investigators saw improvements in cardiac functional indices, indicating that permanent His bundle pacing is a feasible treatment strategy.21 Yet at that time, it never saw widespread adoption as a therapy. Today, His bundle pacing can be achieved by delivering the Select

Secure Model 3830 lead (Medtronic, Minneapolis, MN) through specially designed sheaths (C 315 HIS). His bundle pacing is currently only practiced by a few skilled operators, though there is significant interest in bringing it to the mainstream, as noted above, early studies have shown promising results compared to traditional pacing paradigms. Yet, it should also be noted that several studies have shown mixed results comparing permanent His bundle pacing to traditional right ventricular apical pacing.22–24 Nevertheless, overall acute hemodynamic functional indices appear to improve, when pacing the His bundle, but long-term reductions in heart failure hospitalization and mortality are less clear.25,26

Today, though the basis of evidence to replace traditional right ventricular apical pacing with permanent His bundle pacing is not fully formed, early studies

suggest that pacing the bundle of His may also serve as a replacement for traditional CRT therapy. When compared to traditional left ventricular leads, His bundle pacing has been hypothesized to improve left ventricular functional indices, normalize left ventricular dimensions, and lower NYHA functional class.27,28 Yet currently, there is insufficient data to make definitive conclusions about the safety and efficacy of chronic His bundle pacing. Nevertheless, small early feasibility trials have shown that His bundle pacing may provide significant benefits over other traditional pacing paradigms. These early results, coupled with physiologic intuition make His bundle pacing an attractive frontier for the future of cardiac pacing therapy.

1.11 Future Pacing Directions Pacemakers today are the result of years of innovative ideas combined with extensive clinical expertise. Thus, it follows that the next generation of devices will require more of the same. Our glance here into the future of pacing, shows multiple promising areas of progress that may revolutionize the way we think of pacing today.29

Leadless Pacing: An emerging frontier Many of the next major breakthroughs in cardiac pacing therapies are likely to center around leadless technologies. Leadless pacemakers remove the

Achille’s heal from traditional transvenous systems. The devices that are truly leadless, incorporate pacing circuitry, the battery and the electrodes all within a small intracardiac capsule. Thus, by eliminating lead-related and pocket-related complications, to date, leadless pacemakers have cut implant procedural complication rate to 4%-6.5% from 7.5%-12.5% in traditional single and dual- leaded systems. 30–33 Of note, the procedural complications associated with leadless implants are likely to diminish as the clinicians gain a better understanding of leadless technology and the various implant procedures.

Currently, there are two leadless pacing systems available for clinical use: the Medtronic MicraTM (Medtronic, Minneapolis, MN, USA) and the Abbott

NanostimTM (Nanostim, St. Jude Medical, St. Paul, MN, USA). The Nanostim

LCP received CE Mark in Fall of 2013, and is currently awaiting FDA approval for use in the United States. The Medtronic MicraTM received CE Mark in April 2015, and FDA approval one year later. Both systems are capable of delivery single- 30

chamber right ventricular pacing, sensing, and rate response therapies. These systems are both delivered into the right side of the heart via femoral vein catheterization. A deflectable sheath/delivery tool allows each device to be maneuvered through the right atrium, across the tricuspid valve, and into the desired implant locations within the right ventricle of the heart. However, there are differences in device fixation mechanisms, profiles, deliveries, and functions.

By volume, the MicraTM is the smaller of the two devices (0.8cc vs 1.0cc).

The MicraTM device profile is shorter (25.9mm) with a thicker radial dimension

(23F inner diameter introducer), while the Nanostim is longer (42mm) and thinner radially (18F inner diameter introducer sheath). These devices are significantly smaller than traditional IPG cans. The small size of leadless pacemakers may facilitate the placement of multiple pacing systems within the same heart. In fact, up to three MicraTM pacing systems can be comfortably placed within the right ventricle of the average human heart.34

One of the most critical features of any leadless pacing system will be its fixation mechanism. Leadless pacemakers carry the risk of embolization into the lungs or femoral venous system, if they become dislodged. Interestingly, the two currently clinically available devices utilize fundamentally different fixation mechanisms. The Nanostim has a screw in helix at its distal tip, which anchors the cathode electrode to the ventricular myocardium. In contrast, the MicraTM transcatheter pacemaker fixes to tissue via four self-expanding nitinol tines.

These tines are designed to deploy into the myocardium with minimal tissue

damage, while holding the pacing electrode in stable contact with the ventricular myocardium.35 These differences in fixation mechanism may lead to modest differences in noted pacemaker functionality: both devices utilizations to date have elicited 1.5% pericardial effusion rates in their initial clinical studies.30,31

However, the MicraTM transcatheter pacemaker has shown zero gross dislodgements in clinical trials30, while Nanostim has documented 2.3% dislodgement rate.31

Figure 1.14: Image of a Medtronic MicraTM implanted within the right ventricular apex of a reanimated human heart at the University of Minnesota's Visible Heart Laboratory.

Both of these clinically available leadless pacemakers are designed for rate-responsive ventricular pacing. However, the mechanisms for rate response therapies varies between devices. In the MicraTM pacemaker, a three-axis accelerometer senses extracardiac motions and associatively sets the stimulation rates based on the patient’s relative activity level; in contrast, the

Nanostim device utilizes an incorporated temperature sensor to detect central venous temperatures, and then program an appropriate change in paced rates.36 32

To date, even early in their use histories, leadless pacing systems have shown significant utility, reducing complication rates, while preserving the major functionality of a standalone ventricular pacing system. The devices do, however, face significant questions as to appropriate end-of life decisions. For example, with current battery technologies, these leadless systems are designed to provide approximately 10 years of pacing at stable chronic thresholds

(estimated 12 yrs MicraTM, 8.5-9.8yrs Nanostim). Patients receiving pacemakers are most commonly between the ages of 60-80 years1, but a non-insignificant percentage of patients receive pacing systems much younger. Therefore, the current battery longevities of these leadless systems may necessitate multiple device upgrades and/or revisions, in order to provide long-term therapies to these younger patients. Additionally, should a leadless device become infected, or provide ineffective therapy after implant, the device may need to be physically removed. Indeed, early implantations of the Nanostim pacemaker have shown issues with battery failure, necessitating the removal of many implanted devices.37 As such, appropriate consideration must be given to acute and chronic device retrieval.

Fortunately, recent documented clinical outcomes have supported the acute retrievability of leadless pacemakers.37–39 The Nanostim device has a dedicated retrieval catheter available with either single or tri-loop snares. For the

Nanostim device, successful acute and short-term chronic retrievals were accomplished in 94% of required cases.38 In contrast, the current MicraTM

pacemaker system does not have a dedicated retrieval catheter. To date, retrieval attempts of the MicraTM device commonly involve passing a large diameter snare through a steerable sheath, though a smaller 7mm snare may be used in conjunction with the MicraTM delivery catheter. Direct visualization of leadless pacemaker extraction reveals some of the challenges that face extractors.40 Any extraction attempt must access the proximal retrieval feature of the leadless device; in a chronic setting, the degree of encapsulation of this retrieval feature is unknown. Clinical retrieval experience of the MicraTM leadless pacemaker is limited, but the flexibility of the nitinol tines has facilitated retrieval of the device without counter-traction.39,41 Notably, due to the recent immergence of leadless technology, neither device has proven long-term chronic extractability; the chronic encapsulation profile of leadless devices is yet unquantified.

Current leadless technologies are approved only for right ventricular—

VVI(R)—pacing. To become a truly transformative technology, leadless pacemakers must be adapted for the diverse indications of bradycardia pacing.

Namely, leadless pacemakers should be designed for use in both the atria and ventricles, so to allow for atrial only or dual chamber pacing modalities. For medical device designers, the challenges toward creating leadless dual-chamber systems are considerable. To provide effective therapy, the two leadless devices must be able to decipher what the other is doing. Thus, although direct device- 35

device communication is theoretically possible, it must be accomplished with significant consideration to the finite battery lives of the employed leadless devices. Additionally, designers must account for the substantial differences in tissue biomechanics and hemodynamics across chambers. Simply put, the currently existing ventricular leadless devices are not readily transferable to atrial or dual chamber pacing applications. Therefore, medical device engineers must critically understand clinical needs to deliver therapies that can maximally benefit patient populations indicated for bradycardia therapy.

Endocardial LV Pacing and Novel CRT In addition to bradycardia pacing, the benefits of leadless pacing may be an aid for cardiac resynchronization therapy. Of note, currently, the WiSE-CRT

(WiSE-CRT, EBR Systems, Sunnyvale, California) pacing system utilizes ultrasound technology paired with a leadless endocardial receiving electrode.42

The WiSE-CRT system consists of a traditional dual chamber (RA, RV), and a transmitter/receiver combination that stimulates the left ventricular endocardium.

After a traditional lead applies a pacing stimulus to the right ventricle, an ultrasound wave is emitted by a subcutaneous transmitter. A receiver implanted in the left ventricle converts this ultrasound energy into a pacing stimulus. As such, the system delivers left ventricular synchronization pacing through the endocardial LV. Early studies with this leadless CRT system have shown that the device may be able to improve outcomes in patients who did not respond to traditional CRT therapeutic approaches.42

Conceptually, multiple device solutions, such as those proposed by WiSE-

CRT, can fit a wide variety of clinical needs with these novel technologies. For example, another combination of leadless pacemaker and a subcutaneous defibrillator may be able to substitute for traditional implantable cardio defibrillator leads. Indeed, early studies have aimed to pair the two devices, and suggest that leadless pacemaker performance was not affected by a subcutaneous defibrillation shock.43,44 One limitation of pairing the currently market-available leadless pacemakers with subcutaneous defibrillators is the lack of antitachyarrhythmia pacing capabilities (as can be provided by most implantable cardio defibrillator leads). It should be noted, that Boston Scientific has begun investigation into a new leadless technology that will pair with their market- released subcutaneous defibrillator. In early ovine studies, they have shown feasibility in the design, which would allow device-device communication, and leadless antitachyarrhythmia pacing before attempting defibrillation.45 These up- and-coming technologies may help to substitute for traditional transvenous ICD coils, providing the benefits of leadless technology, while importantly, sacrificing minimal functionality.

Figure 1.16: The WiSE-CRT endocardial LV pacing system. The system utilizes a co-implanted right-sided pacemaker, an ultrasonic transducer, and an endocardial leadless receiver.

The WiSE-CRT system actually combines two common regions of interest for next generation pacing: leadless technology and endocardial LV stimulation.

Currently, all cardiac resynchronization therapy is delivered through epicardial electrodes (generally placed through the venous vasculature. Notably, epicardial left ventricular pacing introduces multiple areas of clinical concern, including: 1) non-response due to unsuitable coronary venous anatomy, 2) phrenic nerve activation, and/or 3) less physiologic epicardial-to-endocardial activation. By

contrast, endocardial left ventricular pacing theoretically allow for pacing from any LV site, perhaps resulting in idealized, more physiologic lead placement.

Additionally, compared to epicardial pacing, it has been reported to improve acute hemodynamic functional indices.46 Further, with endocardial pacing, there is a diminished risk of phrenic nerve stimulation, a more physiologic endocardial- to-epicardial activation pattern, and a larger area of myocardial substrate to pace.

In future pacing therapies, endocardial left ventricular access might gained transapically, transmurally, transaortically, or perhaps across the interatrial or interventricular septae. One must consider that any leaded endocardial left ventricular pacing solutions, one must take care to mitigate the risks of thromboembolism and/or mitral valve regurgitation. Future CRT technologies- whether leaded or leadless—are likely to leverage the benefits of endocardial left ventricular stimulation.

Next Generation Cardiac Resynchronization Therapy: Employing Multisite pacing Importantly, contractile dyssynchrony is the underlying substrate for CRT.

As such, a recent clinical hypothesis postulates that left ventricular pacing at multiple left ventricular locations may improve overall resynchronization, and thus improve both resting cardiac outputs (ejection fractions) and thus clinical outcomes. There are two overarching ideas on how to accomplish this: 1)using multisite pacing with 2 or more LV leads, each placed within the coronary vasculature or 2) employing while multipoint pacing using a single multipolar lead.

Importantly, multisite pacing has been shown to improve acute hemodynamic responses in heart failure patients.47,48 However, there have been noted significant drawbacks to these approaches. Currently, synchronous LV lead timing is accomplished via a Y adapter: thus, the additional hardware leaves multisite pacing systems at a higher risk for chronic loss of capture.49 Battery depletion also remains a significant issue, as the implanted pulse generator is responsible for delivering therapy through two parallelized leads.49 To date, chronic placement of multiple LV leads has yielded mixed results; nevertheless multisite pacing remains an active area of interest.

Quadripolar leads have emerged as a viable means for delivering multipoint pacing. In a quadripolar lead, the clinician attains the ability to choose a pacing vector from four electrodes placed on the distal portions of the lead body (e.g. the pacing vector may be programmed between electrodes 1-2, or 1-4, or 2- 4, etc.). Thus, having the abilty to select the pacing vector, reduces the probability of phrenic nerve capture during pacing, and simultaneously minimizes the chances of high pacing capture threshold (as a separate vector can be selected should capture threshold rise). Early studies show acute hemodynamic advantages and lower mortality rates with quadripolar leads when compared to bipolar leads.50,51

Figure 1.17: Attain Performa quadripolar lead family. Courtesy of Medtronic. Batteryless pacing As noted multiple times within this chapter, battery depletion has been a longtime limiting factor in cardiac pacemaker therapy. Whether considering the first implantable pulse generators, or modern leadless devices, limited battery life remains a significant barrier to advancing pacing therapy. As time has passed, battery chemistry has provided dramatic reductions in battery size (early pacemakers exceeded 300g mass, while current leadless devices sit at less than

2g) while improving battery longevity. An attractive, yet currently clinically unattainable concept is pacing without a battery or with ones that can be recharged transcutaneously.

The human heart beats an estimated 2-3 billion times across an average lifetime. Each beat of the heart couples electrical wave fronts into mechanical beats. As such, any pacemaker energy source must be inexhaustible, if it hopes

to improve upon battery technology. Research efforts have focused on many avenues: perhaps some of the more promising efforts have centered on the beating of the heart to power pacemakers using either mechanical coupling or piezoelectrics.52,53 Other attempts have aimed to harvest solar energy to power cardiac pacemakers.54 These concepts have been tested in animal models, but remain in the realm of research.

Other research efforts have centered around generating biological pacemakers using gene or stem cell therapies.55,56 Biological pacemakers are still early in research phases. There are many challenges to overcome, including, but not limited to engraftment and the significant potential for promoting arrhythmia. Nevertheless, if these problems can be solved they will have applications to all clinically implanted devices within humans (e.g., deep brain stimulators, or LVADs)

1.12 Conclusion This chapter provided an overview of the pacemaker technologies, firmly establishing the vibrant history of the industry, describing state-of-the-art technologies, and providing a glimpse into the future of this bright field.

Throughout the history of the pacemaker, clinical inspiration has driven both ground-breaking research and innovation. The symbiosis between engineering and medicine pivotally shaped the development of these life-saving technologies, and will prove to be just as critical moving forward.

Section II. Anatomic Considerations for Next- Generation Pacemakers

The drive for next-generation pacing systems necessitates a comprehensive understanding of the anatomies in which the devices will be implanted. Importantly, innovation within the pacing field often attempts to expand patient populations and reduce complication rates. This section of my thesis aims to provide anatomic insights into meeting both innovation prongs.

Epicardial Pacing Currently, clinicians and design engineers have begun to work with anatomically innovative solutions which may reduce key complications of traditional pacing systems. Epicardial leads have cycled out of common practice, largely because of the relative difficulty in the surgical implant procedure, less physiologic pacing profile, and higher pacing capture thresholds. However, subcutaneous implantable defibrillators have opened doors to exploring the extracardiac space for novel pacing and defibrillation therapies.57 Notably, the epicardial space has become a target of interest for new therapies aiming to help patients who may not be suited for transvenous pacing.58,59 In the first two chapters of this section, I investigate a key anatomic hurdle that will drive design characteristics for epicardial and extravascular pacing technologies: the epicardial fat pad.

His Bundle Pacing While epicardial pacing therapies aim to reduce complications associated with traditional transvenous leads, they result in a less physiologic paced beat when compared to other pacing therapy. In some corners of the pacing world, the clinicians argue that we ought to strive for the most physiologic pacing capture profile possible; i.e. by pacing into the native His-purkinje conduction system. As previously introduced, His bundle pacing has shown improved acute hemodynamic and electrical cardiac function. Within this section, I look at physiologic pacing through the His bundle and begin to provide key links between His bundle anatomy at the implant site and the resultant QRS complex morphology.

Next-Generation Leadless Pacing Current leadless technologies are approved only for right ventricular—

VVI(R)—pacing. To become a truly transformative technology, leadless pacemakers must be adapted for the diverse indications of bradycardia pacing.

Namely, leadless pacemakers should be designed for use in both the atria and ventricles, so to allow for atrial only or dual chamber pacing modalities. In the final chapter of this section, I look into the key anatomic challenges surrounding the implementation of leadless fixation in the right atrium.

Chapter 2: The Quantitative Assessment of

Epicardial Fat Distributions on Human Hearts:

Implications for Epicardial Electrophysiology

Alexander R. Mattson1,2,3, BS, Mario J Soto3,4, BS, Paul A. Iaizzo PhD, FHRS1,3

1University of Minnesota Departments of Biomedical Engineering and Surgery

2Medtronic, plc

3Institute of Engineering in Medicine, University of Minnesota

4University of Puerto Rico

2.1 Preface

The following chapter highlights a key limiting factor in the design and implantation of epicardial and extravascular pacing systems, namely, the epicardial fat pad.

Often, on the human heart, a significant adipose layer lies superficial to ventricular myocardium, thus preventing efficient electrical interfacing. This chapter seeks to define where epicardial adipose most commonly lies on the heart. It defines common locations where epicardial fat is lacking, deriving probabilistic distributions of where adipose tissues were present across a large number of human hearts. This information is tied back to donor history, providing correlations between patient characteristics and the percent of adipose coverage on their heart.

Importantly, this information is key in defining locations on the heart where epicardial pacing could be performed consistently without interference from adipose tissues. This information is fundamental to a complete understanding of the epicardial anatomy for clinicians hoping to implant in the extravascular or intrapericardial space. Additionally, it provides key insights for device designers hoping to create next-generation pacemakers targeting the epicardial surface of the heart.

Mario Soto contributed to data analysis, manuscript editing, and preparation. Paul Iaizzo contributed to study design and manuscript editing and preparation.

The data in this chapter was partially presented as an oral presentation at the 2017 Biomedical Engineering Society Meeting. The full chapter contents have been published as an original manuscript in Clinical Anatomy:

Mattson AR, Soto MJ, and Iaizzo PA. (2018), The quantitative assessment of epicardial fat distribution on human hearts: Implications for epicardial electrophysiology. Clin. Anat., 31: 661-666.

2.2 Synopsis

Introduction

Epicardial electrophysiological procedures rely on dependable interfacing with the myocardial tissue. For example, epicardial pacing systems must generate sustainable chronic pacing capture, while epicardial ablations must effectively deliver energy to the target hyper-excitable myocytes. The human heart has a significant adipose layer which may impede epicardial procedures. The objective of this study was to quantitatively assess the relative location of epicardial adipose on the human heart, in order to define locations where epicardial therapies might be performed successfully.

Materials and Methods

We studied perfusion-fixed human hearts (n=105) in multiple isolated planes including: left ventricular margin, diaphragmatic surface, and anterior right ventricle. Relative adipose distribution was quantitatively assessed via planar images, using a custom-generated image analysis algorithm.

Results

In these specimens, 76.7±13.8% of the left ventricular margin, 72.7±11.3% of the diaphragmatic surface, and 92.1±8.7% of the anterior right margin were covered with superficial epicardial adipose layers. Percent adipose coverage significantly increased with age (p<0.001) and history of coronary artery disease (p<0.05). No significant relationships were identified between relative percent adipose coverage and gender, body weight or height, BMI, history of hypertension, and/or history of congestive heart failure. Additionally, we describe two-dimensional probability distributions of epicardial adipose coverage for each of the three analysis planes.

Conclusions

In this study, we detail the quantitative assessment and probabilistic mapping of the distribution of superficial epicardial adipose on the adult human heart. These findings have implications relative to performing epicardial procedures and/or designing procedures or tools to successfully perform such treatments.

2.3 Introduction

Epicardial intervention in cardiac electrophysiology is a relatively new frontier: i.e., for performing catheter ablations, alternate site pacing, and/or administering therapeutics. Epicardial procedures typically require a dependable interface with the myocardium; e.g., epicardial cardiac resynchronization therapy

(CRT) relies on stable, chronic pacing capture, while pacing in a location optimized for a given patient. Similarly, epicardial ablations for treating ventricular tachyarrhythmias, require the delivery of sufficient radiofrequency energy to the target myocardium.

A significant adipose layer often lies superficial to the ventricular myocardium, which may impede electrophysiological interventions in the pericardial space. Importantly, adipose tissues elicit poor electrical conductivities. Thus, when present epicardially, they may limit optimal pacing real estate and/or reduce epicardial pacing efficiency. Some epicardial pacing leads have used lengthened, electrically active helices to penetrate through the epicardial adipose layer, in order to gain contact with the ventricular myocardium.

Correspondingly, if one needs to ablate myocardium lying below an epicardial fat layer, then the ablative radiofrequency (RF) energy has to be adjusted accordingly, as applied RF energies are significantly attenuated when delivered through adipose tissue.60 Nevertheless, there is little work describing the potential risks of clinical manipulations of these adipose tissues, either short term or chronically. 49

To date, few studies have attempted to quantitatively assess the relative distributions of adipose on a large population of adult human hearts. In one such study, a reported 80% (56-100%) of the ventricular myocardium was covered with an epicardial fat pad.61 It is commonly accepted that adipose covers the majority of the major coronary vasculature, as well as the atrioventricular groove,62 though distributions relative to specific anatomical locations of the ventricular myocardium remains unclear. Another study suggested that epicardial adipose layers presents four times more on the right ventricle as opposed to the left.63

There have been several investigations relative to possible correlations between epicardial fat mass and a given patient’s age, gender, BMI, and overall heart health, particularly in patients with coronary artery disease.61,62,64 Patient weight and obesity have, in general, been associated with an increase in epicardial adipose.62 Yet, autopsy reports linking BMI with epicardial adipose have proven inconclusive.61,65 Several studies have linked epicardial adipose with the progression of coronary artery disease, with one going so far as to show a increased thickness of adipose tissues on the left ventricular margin associated with increasing severity of coronary artery disease.66 To the author’s knowledge, there is no study which explicitly defines the locations of epicardial adipose.

Hence, with this relative dearth of knowledge surrounding the distribution of epicardial adipose on the human heart, an attempt to quantitatively localize these

tissues across a range of patient histories and demographics could serve to guide clinical electrophysiologic procedures within the pericardial space.

2.4 Objective

In the present study, we examined factors that may correlate with the extent of epicardial adipose coverage, while assessing the global surface areas on the epicardium of a large database of human heart: i.e., identifying where bare myocardium could be accessible. We develop 2-dimensional probability distributions for the relative occurrences of the epicardial adipose coverage on the surfaces of perfusion-fixed human hearts. This library of human hearts associated with varied disease states were investigated to better define possible regions for optimized epicardial pacing and/or epicardial procedures.

2.5 Methods

Human hearts (n=105), were obtained fresh from organ donors whose hearts were considered non-viable for transplant: they were donated for research via our local organ procurement agency (LifeSource, Minneapolis, MN). Relevant donor demographics and clinical diagnoses are presented in table 2.1.

Quantifying Relative Adipose Coverages for Each Specimen

Each donated specimen was cannulated when fresh and then perfusion fixed in a 10% formalin solution, typically less than 24 hours post-cross-clamp 51

time, using previously described methodologies:67 Thus, each specimen was preserved in an approximation of its end-diastolic state. Next, each heart specimen was imaged in each of the following three, 2-dimensional planes: 1) the anterior right ventricle, 2) the diaphragmatic surface, and 3) the left ventricular margin. An in-plane scale bar quantified area and length for each image. A custom-generated algorithm (Matlab, Mathworks, Natick, MA, USA) was written and employed to automate the quantifications of the relative amounts and locations of the epicardial adipose in each stored image.

A B C

LA RA

LAA RA RVOT CS

LV LV RV LV

Figure 2.1: 3-Dimensional reconstruction depicting the epicardial adipose coverage of one donated specimen (HH099, 46 year old male with history of CAD). Images of each donated specimen were taken in three anatomical planes: A--Anterior Right Ventricle, B-- Left Ventricular Margin, C--Diaphragmatic Surface.

Generating Probability Distributions of Adipose Coverage

Sn S2 S1

A B C

Figure 2.2: A—Hearts were imaged in isolated planes, with appropriately positioned scales so to reference dimensions. Regions with and without adipose both were quantified. Common anatomical points selected on the base and apex allow rotation and scaling to align each image on a normalized reference frame. B—Each rotated and scaled specimen image was assigned a weight, according to Equation 2. Weighted images were summed together. C—Weighted sums generated probabilistic distributions of where adipose was/was not on the ventricular surfaces.

Images from each of the 105 heart specimens were scaled, translated,

and rotated based on known anatomical markers, to align all images to a

normalized frame. The final predictions of adipose distribution were created from

a weighted sum of the specimen images. As such, hearts within the dataset that

closely matched input patient characteristics factored more heavily into the

output, while dissimilar hearts factored less into the output.

Notably, only the ventricular adipose layer was examined in this study.

Specimens were donated for research purposes, however other organs,

including the liver and/or lungs, were commonly donated for use in

transplantation. As such, the left and right atrial tissues in the cardiac specimens

were often procured with damage to the posterior aspect. For this reason,

accurate adipose reconstructions for this study were limited to ventricular tissues. 53

Statistical analyses were completed using a series of one-way ANOVA tests, with Tukey honest significant difference post-hoc tests. Comparisons were studied between male and female donors, young (<30 years), middle-aged (31-

60 years), and older (61+ years) donors, and underweight (BMI<25), normal weight (2530) donors. Populations of specimens eliciting various disease states were compared to a cohort of control hearts

(those without relevant disease history).

2.6 Results

Of the 105 perfusion fixed human heart studied, there was a similar distribution of male and female specimens, with a slight weight toward female specimens. The average donor age was 54.3 years, with a range of 14-82 (Table

1). Of these specimens, 19 hearts were classified as eliciting coronary artery disease (CAD), 33 with hypertension (HTN), 9 with congestive heart failure and

57 were considered as without known cardiovascular disease. With respect to

BMI, patients were roughly evenly distributed, with approximately one third of donations originating from patients with BMI <25, one third patients with

2530.

Table 2.1: Donor Patient Demographics/Characteristics and Relation to Adipose Coverage Characteristic Number of Anterior RV LV Margin Diaphragmatic Patients Adipose % Adipose % Surface (p-value) (p-value) Adipose % (p-value) Gender (p = 0.40) (p = 0.66) (p = 0.53) Male 50 (47.6%) 95.1 77.4 73.4 Female 55 (52.4%) 90.9 76.1 72.0 Age µ = 54.3 ± 14.5 (p < 0.001) (p < 0.001) (p < 0.001) 0-30 years 12 (11.4%) 77.7 59.1 60.7 31-60 years 60 (57.1%) 92.2 75.3 72.0 61+ years 33 (31.4%) 94.8 83.8 78.0

BMI µ = 27.5 ± 7.6 (p= 0.50) (p = 0.97) (p=0.90) Unknown 5 (4.8%) <25 (Normal) 35 (33.3% 90.8 76.2 72.2 25-30 (Overweight) 34 (32.4%) 91.4 77.1 72.6 >30 (Obese) 31 (29.5%) 93.3 76.7 73.5 Disease State No Relevant History 57 (54.2%) 90.1 74.7 70.9 Hypertension 33 (31.4%) 96.2 (p = 0.003) 79.7 (p = 0.09) 73.5 (p = 0.23) Diabetes 21 (20.0%) 94.6 (p = 0.10) 77.3 (p = 0.52) 74.4 (p = 0.21) Congestive Heart Failure 9 (8.6%) 91.9 (p = 0.66) 71.1 (p = 0.50) 73.5 (p = 0.50) Coronary Artery Disease 19 (18.1%) 94.9 (p=0.03) 78.5 (p = 0.29) 75.3 (p = 0.09)

Within healthy control specimens, superficial epicardial adipose tissue covered an average of 74.7±13.8% of the left ventricular margins, 70.9±11.3% of the diaphragmatic surfaces, and 90.1±8.7% of the anterior right margins. The relative percentage of adipose tissue coverage on the hearts increased globally with age (p<0.001). On the anterior right ventricle, history of hypertension and history of coronary artery disease correlated with greater adipose coverage. In

this analysis, gender, heights, and weights were not significant factors driving adipose coverage. Interestingly, BMI was the weakest predictor of eliciting a given degree of adipose coverage.

Despite substantial adipose coverage on both the right and left ventricles, derived probability distributions show that several cardiac regions exhibit low probabilities of adipose coverage. Probability distributions of epicardial adipose for the three analytical planes can be seen in Figures 2.3-2.5. The pictured probability distributions were modeled for a 60 year-old male with no relevant cardiac history. Bare surfaces on the anterior right ventricles, typically take “L” shapes: i.e., following the contours of the right ventricular outflow tracts. On the model of the average human heart, this region is comprised of 6.2±6.5cm2 of bare cardiac myocardium. On the modeled diaphragmatic surface, it consists of large, elliptical patches present on both the right and left ventricles: i.e., on either side of the interventricular septum. The exposed myocardial areas on the diaphragmatic surface averaged 12.0±7.3cm2 on the LV and 14.7±7.8cm2 on the

RV. The derived probability of adipose distribution was more diffuse on the left ventricular margin. The average human heart, had 14.0±10.2cm2 bare area on the left ventricular margin: centered on the basolateral mid-wall directly under the left atrial appendage and extending to the posterolateral surface of the LV.

Figure 2.5: There is low probability of adipose coverage on the modeled right and left ventricle on either diaphragmatic surface. These exposed myocardial patches were the most consistently present across this series of human hearts: as was evidenced by the extremely low calculated probability of epicardial adipose in this region.

2.7 Discussion

In the present study, we investigated the relative distributions of adipose on the epicardial surfaces of human hearts (n=105). To the author’s knowledge, the present study is the first to probabilistically assess the coverage of epicardial adipose and correlate these distributions relative to patient demographics.

Comparison to Previous Literature

Here we observed that the left ventricles of the human hearts typically possess significantly less adipose coverage than the right ventricles, which is consistent with previous studies showing right ventricular adipose mass exceeding that of the left ventricle.68 Age has previously been linked with increases in epicardial adipose mass and thickness.62 Consistent with those observations, we determined that a greater percentage of ventricular myocardium was covered by epicardial adipose in the heart specimens we studied which were obtained from older organ donors. Likewise, as was previously reported that epicardial adipose mass plateaus at age 40 years62, we also observed attenuated increases in adipose coverage when comparing older patients .

The presence of epicardial adipose has been described to play a driving role in the progression of coronary artery disease.69 Further, it was previously noted that patients with increased epicardial adipose volume elicited increased 58

severities of coronary plaques.62 Here, we show that a history of coronary artery disease corresponded to a statistically significantly greater surface area of right ventricular myocardium being covered by adipose tissue: suggesting that epicardial adipose may grow both in thickness and in coverage with the progression of coronary artery disease.

Unlike previous analyses, in the present study we did not find a significant relationship between BMI and the relative degree of surface area coverage of adipose. Previously, BMI has been linked to increases in enhances adipose volumes, masses, and thicknesses.62,68 Here, despite a significant range of donor

BMI, 105 specimens with BMI ranging from 15 – 51 kg/m2, we found no statistical correlation between surface area of adipose coverage and BMI. In fact, in this analysis, BMI was one of the weakest predictors of percent adipose coverage.

This may suggest that increases in BMI increase adipose thickness, while minimally impacting adipose distribution; i.e. more adipose does not necessarily correspond to greater surface area coverage of adipose: additional investigations will be required to validate such.

Implications for Device Therapy

We note that the presented results may have lasting implications for the clinical applications of epicardial pacing, defibrillation, and/or ablative therapies.

In other words, the quantitative, probabilistic roadmap of epicardial adipose may help guide interventions the require applications within the pericardial space.

The optimization of cardiac resynchronization therapy is highly reliant on achieving left ventricular pacing capture in a patient-specific restorative region.

Previous studies have postulated that the posterolateral region of the left ventricle, often the most delayed region in left ventricular systole, frequently offers an anatomical site for this ideal pacing substrate.70 Further, epicardial lead placement has been suggested to increase CRT response rates in those that do not respond to traditional transvenous CRT.71,72

Here, we reported that the posterolateral region of the left ventricle had a low probability of adipose coverage, suggesting that the average human heart may have exposed myocardium in a clinically appropriate region for the application epicardial pacing leads. In other words, this approach may open the door for CRT in patients whose optimal pacing location could not be achieved via percutaneous CRT lead placement.

With the impending rise of extravascular pacing and defibrillation technologies, an accurate characterization of epicardial fat is essential in defining unimpeded myocardial substrate. In this study, we found that the right ventricular outflow tract in the studied cardiac specimens commonly elicited an “L-shaped” region of minimal adipose coverage, offering a reasonable target for extravascular pacing and/or defibrillation leads to be placed within the anterior mediastinal space. Further, we suggest that placing an extravascular lead too close to the right ventricular margin or the interventricular groove will likely result

in attempting to pace or electrically sense through a region with higher adipose probability, likely reducing efficient electrical interfacing.

Our presented research may also aid in the planning of epicardial ablation procedures, which recently has become a useful tool to treat a variety of ventricular tachyarrhythmias associated with various cardiomyopathies.

Radiofrequency ablative energy is known to be significantly attenuated when applied through epicardial adipose, potentially reducing efficacy of the targeted ablation.73 Importantly, previous studies attempting standard RF epicardial ablations were unable to produce lesions through adipose layers greater than

3.1mm thick60, illustrating the challenges faced when ablating targets on the ventricular epicardium. The present study provides a quantitative localization of regions with high and low epicardial fat probability, identifying regions where epicardial ablation may or may not be a suitable treatment modality.

2.8 Conclusion

Here, we have detailed a quantitative 2-dimensional computational model of the relative distributions of epicardial adipose on the ventricular epicardium of human hearts. The presentation of epicardial adipose coverage followed characteristic patterns, leaving regions of high and low probability of exposed myocardium. This probabilistic mapping of epicardial adipose is considered important so to best define the relative myocardial anatomies available for

several electrophysiological procedures, including, but not limited to epicardial pacing, epicardial ablation, and/or extravascular pacing and defibrillation.

Chapter 3: 3-Dimensional Reconstruction and

Quantitative Assessment of Ventricular Epicardial Fat

Tissue on Human Hearts

Alexander R. Mattson, BS1,2,3, Susan Y. Sun, MS1, Traci Jones, MS1,2, Paul A. Iaizzo PhD, FHRS1,3

1Departments of Biomedical Engineering and Surgery, University of Minnesota,

Minneapolis, MN USA

2Medtronic, Mounds View, MN USA

3 Institute for Engineering in Medicine, University of Minnesota, Minneapolis, MN

USA

3.1 Preface

The previous chapter defined the most common locations of epicardial adipose on the human heart. This chapter dives further, looking into the thickness of epicardial adipose at each of these locations.

Pacemakers may be able to deliver pacing therapy through a thin epicardial adipose layer, but their efficacies are significantly diminished when delivered through thick adipose. Further, when attempting to place epicardial leads, surgeons must often tunnel through adipose tissue to reach viable myocardium.

This chapter identifies regions of the heart where adipose is thickest, defines patient characteristics driving adipose thickness, and introduces a library of 3-dimensional adipose reconstructions, generated from MRI scans of donor hearts.

The information in this chapter is of critical importance to any clinician performing epicardial ablation and/or pacing, in addition to design engineers hoping to create pacing solutions within the pericardial or extravascular spaces.

The data analysis and experimental design in this chapter were completed collaboratively by me, Susan Sun, and Traci Jones. Paul Iaizzo, Susan Sun, and

Traci Jones contributed to preparation and editing of the following chapter.

This chapter was submitted as an original manuscript for publication in Europace.

3.2 Synopsis

Introduction

Epicardial interventions have forged new frontiers relative to cardiac catheter ablation, and device therapy. Yet, on the typical patient’s heart, a significant adipose tissue layer lies superficial to the ventricular myocardium; which may hinder success and/or increase complexity of epicardial interventions. This chapter quantitatively evaluates the distribution of epicardial adipose tissue on the surface of human hearts, and provides high fidelity, 3-dimensional reconstructions of the observed epicardial adipose tissues.

Methods

Human hearts (n=80), deemed non-viable for transplant were donated for research via our local procurement agency (LifeSource, Minneapolis, MN). Each donor heart was perfusion-fixed in a 10% formalin solution, within 24 hours of cross-clamp. Subsequently, donor hearts were MRI scanned in a 3T machine using an mprage (T1 weighted) protocol. MRI scans were used to reconstruct 3- dimensional anatomical representations of the donor anatomies using Mimics software (Materialise, Leuven Belgium). The regional thickness of the adipose tissue was analyzed at 51 anatomical reference points surrounding both the left and right ventricles. Adipose thickness was compared to a given patient demographic.

Results

Adipose deposits on the human hearts displayed characteristic patterns, with thick accumulations along the interventricular septa (anterior—9.01±0.50mm; posterior—6.78±0.50mm) and the right ventricular margin

(7.44±0.57mm).Conversely, adipose thickness was minimized along the diaphragmatic surface (RV—1.14±0.82mm; LV—2.02±0.28mm), left ventricular free wall (1.40±0.46mm), and right ventricular outflow tract (3.24±0.64mm).

Conclusions

We provide here, one of the most complete characterizations of human heart epicardial adipose locations and thicknesses to date. These results are fundamental for an underlying anatomic understanding when performing electrophysiologic procedures in the pericardial space.

3.3 Introduction

The use of the intra-pericardial space for both electrophysiologic mapping and ablative procedures was not fully appreciated by cardiac electrophysiologists, until the turn of the century, when the first case reports of epicardial catheter ablation for ventricular tachyarrhythmia were published.74 . What was once disregarded in the clinical world, has recently become a new frontier in cardiac catheter ablation, and device therapy. Importantly, epicardial intervention typically requires a dependable interface with the ventricular myocardium. On the human heart, a significant adipose layer often lies superficial to the ventricular myocardium, which may hinder success and increase complexity of attempted epicardial interventions. For example, the poor conductivity of adipose tissue, will limit real estate for the pacing lead to contact myocardium and/or impede efficient delivery of the pacing stimulus. Additionally, applied radiofrequency energies are significantly attenuated when delivered through adipose tissue.60 Correspondingly, radiofrequency ablations of myocardium lying below an epicardial fat layer, must tune the delivered energy to propagate through the adipose tissue, while minimizing excessive damage to surrounding tissue (e.g. coronary vasculature).

Few reported studies have attempted to quantitatively assess the relative thickness distribution of adipose tissue on large populations of human hearts.

Yet, the basic overall profile of epicardial adipose has been described: e.g.. an estimated 80% (56-100%) of the ventricular myocardium is covered with an 67

epicardial fat pad.61 Further, previous studies have shown that adipose tissue covers the majority of the major coronary vasculature, as well as the atrioventricular groove.62 Distributions relative to specific locations of the ventricular myocardium remain unclear. In one study it was suggested that a greater amount of adipose coverage, in both mass and percent coverage, occurs on the right ventricle, when compared to the left.63 Several additional reports describe that epicardial adipose may correlate with a given patient’s age, gender,

BMI, and/or overall heart health/weight.61,62,64 Our laboratory previously identified regions of the ventricular myocardium which were most commonly covered by a superficial adipose layer.75 Nevertheless, although there have been numerous reports describing the qualitative distributions of epicardial adipose on human hearts, there has been no precise quantitative characterization of epicardial adipose distributions and their relative thicknesses as a function of anatomic location on the heart.

Our investigation quantitatively evaluated the distribution of epicardial adipose thicknesses for a large sampling of 80 perfusion fixed human hearts. We have provided high fidelity, 3-dimensional reconstructions of these epicardial adipose tissues, and critically examined relative patient characteristics which may correlate with various patterns in adipose coverage.

3.4 Methods

Human hearts (n=80), deemed non-viable for transplant were donated for research to the Visible Heart® laboratory via our local procurement agency

(LifeSource, Minneapolis, MN). A summation of donor characteristics can be found in Table 3.2: additional donor information on these pictured specimens can be found on the free-access website, “The Atlas of Human Cardiac

Anatomy”, www.vhlab.umn.edu/atlas/histories/histories.html) .

Each donor heart was perfusion-fixed in a 10% formalin solution, within 24 hours of organ recovery. Subsequently, the hearts, saturated with formalin, were placed in sealable polymer containers and embedded in 7% agar gel to stabilize the specimens for imaging. These gelled and fixed donor hearts were MRI scanned using in a 3T machine (Siemens TRIO, Siemens Corp., Washington

DC, USA) using an MP-RAGE (T1 weighted) protocol, with base resolution of

512 voxels. Scan slice thickness varied between 0.8-1.2 mm; i.e., dependent on the relative size of the donor heart. Each scan view was oriented to obtain axial slices of the imaged specimen (Figure 3.1A).

A B

Figure 3.1: (A) Shown here is a typical example of an axial slice of a high-resolution MRI scan of an isolated human heart. Regions with adipose tissue lying superficial to the ventricular myocardium are indicated with white arrows. (B) 50 anatomical reference points encircling both the left and right ventricles were selected. At each point, the relative adipose thickness was measured. Subsequently, the thicknesses at each defined anatomic location were averaged across of the entire sample of human hearts. TV = Tricuspid Valve, MV = Mitral Valve, PV = Pulmonary Valve, AIS = Anterior Interventricular Septum, PIS = Posterior Interventricular Septum

Each high-resolution cardiac MRI DICOM image set was subsequently reconstructed into 3-dimensional anatomical representations of the donor heart anatomies using Mimics software (Materialise, Leuven Belgium). Segmentations of epicardial adipose distributions were generated, to create full anatomically- accurate reconstructions of the ventricular myocardium and the layers of superficial adipose tissue (Figure 3.2).

Notably, only the ventricular adipose layer was quantitatively examined in this study. These heart specimens were donated specifically for research purposes; however, from some of these donors, other organs typically including the liver and/or lungs, were donated for clinical transplantations. As such, the left and right atrial tissues in these cardiac research specimens, were often damaged. For this reason, accurate adipose reconstructions for the present study, were limited to ventricular tissues.

In addition to full adipose reconstructions for each individual heart, the focal thickness of the adipose layer was analyzed at 51 anatomical reference points surrounding both the left and right ventricles: i.e., so to generate mean values for adipose thickness across many donor specimens. Statistical comparisons across anatomical locations were completed using one-way analyses of variance with a Tukey honest significant difference post-hoc tests.

Adipose thickness distributions were compared to patient demographics, as available from clinical donor records. The typical demographics available for these analyses included: age, gender, height, weight, BMI, and race.

Additionally, any prior diagnoses of a given donor disease state were compared to resultant adipose coverage. Disease states analyzed included: 1) diabetes; 2) history of hypertension (>5 years since initial diagnosis) 3) history of coronary artery disease (defined as previous stent placement, coronary artery bypass graft, >30% occlusion in major coronary vessel precluding donor heart transplant and/or listed clinical diagnosis of coronary artery disease); 4) atrial fibrillation (>1-

year post-diagnosis); and 5) congestive heart failure. Statistical comparisons across patient demographics were completed with an N-way analysis of variance to determine if the mean adipose thicknesses of a given patient subset, differed with respect to groups of multiple patient demographics.

3.5 Results Table 3.1: Donor Patient Histories for Hearts Depicted in Figure 3.3-3.5 Heart Gender Age BMI Weight Height Cardiac Medical Systemic Medical Number (kg/m2) (kg) (cm) History History

084 M 36 33.5 103.0 175 Moderate LVH, LAE, History of smoking dilated IVC, and alcoholism. hypertension, cardiac arrhythmia.

097 F 68 24.7 65.5 163 Hypertension None known

099 M 46 33.8 100.0 172 Severe RCA CAD, Blood clotting drug-eluting stent in disorder, diverticulitis, RCA, severe LV history of smoking (34 dysfunction, years), type II diabetes, hypertension. sleep apnea.

131 M 51 32.8 109.9 183 Hypertension. Diabetes, chronic kidney disease, hyperlipidemia.

140 F 78 29.0 72.0 158 None known. History of smoking (60 years), hypothyroidism.

141 F 24 31.2 86.9 167 None known. None known.

165 F 62 28.5 78.4 166 Hypertension (3 yrs). Diabetes insipidus, history of smoking (5 yrs), hyperthyroidism, hyperlipidemia.

198 F 52 26.1 71.0 165 MI (1 yr prior), stent Type I diabetes (1 yr prior), CAD mellitus, history of smoking (30 yrs), Grave's Disease.

229 F 44 31.5 83.8 163 Hypertension. History of smoking (28 yrs), microcytic anemia, Grave's Disease, hyperthyroidism.

251 M 58 31.8 97.3 175 Chronic Heart Histoplasmosis, Failure, History of smoking, Cardiomyopathy, type II Diabetes, AVNRT, Defibrillator interstitial lung Placed 5 years prior disease, COPD.

Full Epicardial Adipose Reconstructions Full three-dimensional adipose reconstructions from 10 human hearts are depicted in figures 3.3 through 3.5: notably, adipose depositions vary between different hearts. However, despite wide anatomic variability, adipose deposits were elicited in broadly similar patterns across these hearts (Figure 3.3,Figure

3.4, and Figure 3.5). Adipose thicknesses were maximal at the interventricular grooves, the atrioventricular grooves, and the right ventricular margins.

Typically, there were two identifiable elliptical regions of thin adipose and/or bare myocardium which sat on either side of the posterior interventricular septum

(Figure 3.4). These elliptical patches were the most consistent thin/bare region observed from heart-to-heart; nearly every heart elicited some exposed myocardium in this region. Additionally, on the anterior right ventricle, there was often an L-shaped region, which exhibited low adipose thickness, spanning from the right ventricular apex up along the length of the right ventricular outflow tract

(Figure 3.3). Left ventricular adipose deposits were in general thinner, and less consistently-located with respect to major cardiac landmarks (Figure 3.5). In all, the left ventricle showed significantly thinner and more dispersed adipose deposits when compared to the right ventricle.

Figure 3.3: 3-Dimensional adipose reconstructions of the anterior right ventricles from 10 donated human hearts are shown. Here, the thickness of the superficial adipose layer is color-mapped, with thicker to thinner adipose regions indicated by a relative red-to-green color scale (depicted right, in mm); regions devoid of adipose are colored grey. Adipose deposits were thickest at the anterior interventricular borders and at the basal atrial-ventricular grooves. A majority of the anterior right ventricles were covered with at least a thin layer of adipose. The thinnest regions of adipose on the anterior right ventricles followed an “L-Shaped” pattern; typically extending from the right ventricular apex, to the body of the right ventricle, through the right ventricular outflow tract, terminating at the base of the pulmonary artery. Specimens from left to right: (Top) Heart 099, Heart 097, Heart 131, Heart 141, Heart 198, (Bottom) Heart 165, Heart 140, Heart 084, Heart 251, Heart 229. Patient histories are available at the University of Minnesota’s Atlas of Human Cardiac Anatomy free access website (vhlab.umn.edu/atlas/histories/histories.shtml).

elicited two elliptical regions flanking either side of the posterior interventricular septum which showed either thin adipose, or exposed myocardium. These two elliptical regions were the most consistent exposed/thin adipose regions from heart-to-heart, with the ellipses on the right ventricles generally being slightly larger than that of the left. Specimens from left to right: (Top) Heart 099, Heart 097, Heart 131, Heart 141, Heart 198, (Bottom) Heart 165, Heart 140, Heart 084, Heart 251, Heart 229. Patient histories are available at the University of Minnesota’s Atlas of Human Cardiac Anatomy free access website (vhlab.umn.edu/atlas/histories/histories.shtml).

Figure 3.5: 3-Dimensional adipose reconstructions of the left ventricular margins from 10 donated human hearts are shown. Here, the thicknesses of the superficial adipose layers were again color-mapped, with thicker to thinner adipose regions indicated by a red-to-green color scale (depicted right, in mm); regions devoid of adipose are colored grey. Adipose is thinner and scarcer on the left ventricular surface. Typically, the thinnest regions of adipose coverage were located along the posterior lateral wall of the left ventricle. Adipose distributions were thinner toward the base of the left ventricle (below the atrial-ventricular groove), and thicker toward the apex. Specimens from left to right: (Top) Heart 099, Heart 097, Heart 131, Heart 141, Heart 198, (Bottom) Heart 165, Heart 140, Heart 084, Heart 251, Heart 229. Patient histories are available at the University of Minnesota’s Atlas of Human Cardiac Anatomy free access website (vhlab.umn.edu/atlas/histories/histories.shtml).

Defined Areas for Adipose Thickness Measurements

Figure 3.6: Mean adipose thickness determined for the right ventricle across the evaluated 80 human hearts. Values re expressed in millimeters of adipose thickness; the color of each cell in the thickness map (left) corresponds to the mean thickness displayed, with red indicating thicker adipose coverage and green indicating thinner adipose coverage. Regions of interest on the right ventricle may be mapped between the anatomic reconstruction (right Heart 229) and the thickness map (left). Circled in green, was an elliptical zone on the diaphragmatic inferior surface of the right ventricle, which commonly showed the lowest average adipose thickness. Circled in orange, is a region on the right ventricular margin which commonly had the thickest adipose deposit on a given the right ventricle. Circled in black, is a zone on the right ventricular outflow tract; this region typically had the thinnest adipose layer on the anterior portion of the right ventricle. RVE = Right ventricular ellipse, RVM = Right ventricular margin, RVOT = Right ventricular outflow tract.

Average adipose thicknesses across the 80 analyzed human hearts are depicted in Figure 3.6 and Figure 3.7. Additionally, the individual anatomic regions of interest are outlined in Figure 3.6 and Figure 3.7, with mean quantification provided in Figure 3.8. In all, adipose coverage was shown to be the thickest along the posterior and interventricular septa, as well as on the right 78

ventricular margin. Adipose coverage was significantly thicker on the right ventricle, as compared to the left (RV Overall > LV Overall, p < 0.001). The regions of thinnest adipose depositions on the left ventricle were elliptical regions on the posterior diaphragmatic surface (LV Ellipse < LV Overall, p = 0.003) and the posterior left ventricular free wall (LV FW < LV Overall, p = 0.02). On the right ventricle, the right ventricular outflow tract and posterior diaphragmatic ellipse elicited the thinnest average adipose deposits (RVOT < RV Overall, p<0.001; RV Ellipse < RV Overall, p<0.001; RV Ellipse < RVOT p<0.001). The right ventricular margin displayed significantly thicker adipose coverage than the right ventricle as a whole (RV Margin > RV Overall, p<0.001).

septa. Anatomic subsections, as quantified here, were defined in Figure 3.6 and Figure 3.7. Error bars represent the standard error of the mean. RV Ellipse = 1.14±0.82mm, LV Ellipse = 1.27±0.64mm, LVFW = 1.40±0.46mm, LV Overall = 2.02±0.28mm, RVOT = 3.24±0.64mm, RV Overall = 4.73±0.28mm, PIS = 6.78±0.50mm, RV Margin = 7.44±0.57mm, AIS = 9.01±0.50mm. RV = Right Ventricle, LV = Left ventricle, LVFW = Left ventricular free wall, RVOT = Right ventricular outflow tract, PIS = Posterior interventricular septum, AIS = anterior interventricular septum.

Donor Characteristics in Relation to Adipose Thickness Adipose thicknesses were characterized in relation to various donor cardiac histories/characteristics, to ascertain correlations between patient demographics/disease states and adipose thickness (Table 3.2). Adipose thicknesses within defined anatomical areas of the left and right ventricle are compared against patient demographics in Table 3.2. All in all, the best predictor for increased adipose coverage was increased donor age (p<0.001). Adipose thicknesses on both the left and right ventricles increased with increases in age, approximately doubling between young (0-40 years) and older (61+) age groups

(95% Confidence interval for β of regression of mean adipose thickness—mm onto Age—years was [0.04, 0.10] for the RV and [0.02, 0.06] for LV).

Additionally, left ventricular adipose was found to be statistically thicker in female donors as compared to their male counterparts (p=0.01); this relationship did not translate to the right ventricle. There were no patient disease states which correlated significantly with adipose thicknesses, either overall, or within ventricular subsections; here, tested disease states included: congestive heart failure, coronary artery disease, diabetes, atrial fibrillation, and a history of 5+ years of hypertension. Interestingly, there were no statistical differences in 80

adipose thicknesses across BMI; in fact, there was a slight negative correlation between BMI and adipose thicknesses on both the right and left ventricles (95%

Confidence interval for β of regression of mean adipose thickness—mm onto

BMI—kg/m2 was [-0.10, 0.06] for RV and [-0.06, 0.04] for LV).

Table 3.2: Donor Patient Demographics/Characteristics and Relation to Overall Mean RV and LV Adipose Thickness

Mean RV Mean LV Characteristic Number of Adipose Adipose Patients Thickness mm Thickness mm (p-value) (p-value) Gender (p = 0.98) (p = 0.03) Male 37 (46.4%) 4.4 1.7 Female 43 (53.6%) 4.7 2.5 Age µ = 58.1 ± 13.9 (p < 0.001) (p < 0.001) 0-40 years 9 (11.3%) 2.8 1.0 41-60 years 28 (35.0%) 4.0 2.0 61+ years 43 (54.0%) 5.3 2.4 BMI (kg/m2) µ = 27.2 ± 5.7 (p= 0.55) (p = 0.77) <25 (Underweight/Normal) 35 (33.3%) 4.8 2.3 25-30 (Overweight) 34 (32.4%) 4.5 2.2 >30 (Obese) 31 (29.5%) 4.3 1.8 Disease State No Relevant History (Control) 19 (23.8%) 4.5 2.1 Hypertension 36 (45.0%) 4.4 (p = 0.06) 2.1 (p = 0.35) Diabetes 15 (18.8%) 5.0 (p = 0.13) 2.3 (p = 0.20) Congestive Heart Failure 9 (11.3%) 4.8 (p = 0.24) 2.2 (p = 0.43) Coronary Artery Disease 19 (23.8%) 4.8 (p = 0.33) 2.4 (p = 0.23) Atrial Fibrillation 14 (17.5%) 5.3 (p = 0.28) 2.7 (p = 0.70) COPD 6 (7.5%) 4.8 (p = 0.73) 1.7 (p = 0.43)

3.6 Discussion To our knowledge, this study comprises the most detailed, focal anatomic characterization of epicardial adipose tissue distributions on adult human hearts.

We have created detailed 3-Dimensional models for the epicardial fat distribution of 80 perfusion-fixed human hearts from high resolution MRI scans.

Subsequently, we provided precise descriptions of regional adipose thickness variation and described how patient demographics impact distributions. This unique presentation of characterized anatomy should provide a critical component for understanding electrophysiologic interventions to be performed within the pericardial space.

Adipose Coverage Variation by Anatomic Location Adipose is well known to cover the major coronary vasculature.62 In our analysis, the thickest deposits of adipose identified do, indeed, cover the major coronary vasculature including: 1) together the left anterior descending coronary artery and anterior interventricular vein, 2) together the posterior descending artery and middle cardiac vein, and 3) the right marginal artery. Notably, adipose thickness appeared to decrease as a function of vessel diameter; i.e., moving from base to apex along both the anterior and posterior interventricular septa.

This trend, however, does not hold true along the right ventricular margin, where adipose thickness remained relatively consistent from base to apex. Additionally, despite a relatively small average vessel diameter along the right ventricular margin, this region elicited significantly thicker adipose than all analyzed regions

except the anterior and posterior interventricular septa. As such, while vessel diameter appears to play role in driving adipose thickness, there must be other major variables at play.

Previous studies have suggested a greater amount of adipose mass on the right ventricle, when compared to the left.63 Here in agreement, we observed that overall, the right ventricle presents with significantly thicker adipose than the left ventricle (p<0.001). In addition to an overall thicker adipose bed, the right ventricle also has fewer regions with adipose thickness less than 2mm. This corresponds to our qualitative observation that the left ventricle had significantly more surface area devoid of adipose; a phenomenon which can be easily observed when contrasting Figure 3.3 with Figure 3.5. Despite thicker overall adipose coverages, the right ventricle did show two consistent regions where adipose coverage was minimized: an “L-shaped” region on the right ventricular outflow tract, and an ellipse on the posterior diaphragmatic surface of the right ventricle. These two ellipses, mirrored on either side of the posterior interventricular septum were the two regions with the thinnest adipose on the heart (p<0.001). Further, these two elliptical regions were the most consistent regions with exposed myocardium in each heart.

Patient Demographics and Relation to Adipose Coverage Previous investigators have attempted to establish a link between adipose depositions and patient demographics and/or disease states. One of the

strongest reported correlations associated with cardiac adipose coverage described both increases in adipose mass and percentage of adipose coverage with increasing age.64,68,76 This result was confirmed in our study, as we find that epicardial adipose thickness statistically increased as a function of patient age

(p<0.001), doubling in thickness between the youngest and oldest patient groups analyzed. Here, we also observed that the mean epicardial adipose thicknesses on the left ventricle were significantly higher for female donor hearts than for males (p = 0.01).

There have been various studies reporting links between the progression and severity of coronary artery disease with increased epicardial adipose.61,77–79

Taken together, these groups of investigators agreed that a positive correlation existed between epicardial adiposity and progression of coronary artery disease.

To the contrary, in the collection of 80 donor hearts, we observed no relationship between coronary artery disease and epicardial adipose thickness, either globally or regionally (notably, this includes the anterior and posterior interventricular septa and right ventricular margin, where major coronary arteries are present).

However, in this study, we had only information on presence, but not severity or clinical course of coronary artery disease in these patients.

Previous anatomical investigations have reported an increase in epicardial adipose volume, peri-coronary adipose thickness, and right ventricular adipose thickness on the hearts of obese patients (BMI>30 kg/m2).80 Here we observed no significant relationship between BMI and adipose thickness. Yet, overall left

and right ventricular adipose thicknesses trended downward with increasing BMI.

Further, none of the quantitative assessments for the individualized anatomic locations in our study provided significant changes in adipose thickness with increasing BMI.

Study Implications for Cardiac Epicardial Electrophysiology We consider, that the results discussed here have important implications for performing various clinical interventions within the pericardial space.

Epicardial pacing, defibrillation, and/or ablative therapies are all thought to require reliable interfacing with ventricular myocardium. A potentially significant challenge to success of these intra-pericardial interventions is presence of epicardial adipose tissue.

Recently, epicardial ablation of the right ventricular outflow tract has become an attractive therapy to treat patients with ventricular tachyarrhythmia.

However, radiofrequency energy delivery is significantly attenuated by the presence of epicardial adipose layers.81 For example, one study suggests no appreciable lesion formation through ~3mm of adipose when using standard radiofrequency ablation protocols; although, significant, but attenuated lesions were seen when RF therapy was delivered via a cooled-tip radiofrequency ablation catheter.60 As noted above, the right ventricular outflow tract is a common target for epicardial ablation for treating ventricular tachyarrhythmia.

Fortunately, on most of the hearts studied here, the right ventricular outflow tract elicited reduced adipose thickness compared to the right ventricle as a whole.

Nevertheless, in general, adipose thickness averaged over 3mm across nearly the entire anterior surface of the right ventricle (Figure 3.6). As such, standard radiofrequency energy delivery may not be effective in consistently generating effective lesions. Perhaps other ablative modalities, e.g. cryothermal, chemical, electroporative, may have differing efficacies when compared to radiofrequency for treating the underlying myocardium through a fat layer. Nevertheless, future experimentation will be needed to identify whether or not these modalities can improve on the therapeutic efficacies of radiofrequency energies; i.e., given the anatomic challenges of the intrapericardial space.

One also needs to consider that thick regions of adipose will likely attenuate sensed electrogram amplitudes during epicardial voltage mapping. As such, regions with thicker adipose coverage may even resemble myocardial scar tissue on a voltage map.82,83 The novel quantifications of epicardial adipose presented here provide detailed maps of thickness localizations. This data may suggest that epicardial voltage mapping of the right ventricle may have reduced accuracy in certain patients; i.e. the right ventricular margin, anterior/posterior septa, and right ventricular apex all show substantial adipose thicknesses that may confound mapping results.

The results presented here also have implications for the successful delivery of epicardial pacing therapy. In general, optimized cardiac resynchronization therapy relies on pacing capture in patient-specific restorative regions. The posterolateral region of the left ventricle, commonly the most

delayed region of left ventricular activation, is often the ideal pacing substrate/location.70 Here, we observed that the posterolateral left ventricle elicited some of the lowest average adipose thicknesses on the entire ventricular myocardium (Figure 3.7). A significant portion of our study hearts would have had readily accessible myocardium for therapeutic pacing; i.e., unimpeded by adipose along the lateral left ventricular wall (Figure 3.5). Thus, the adult human heart may commonly present anatomically favorable locations for placement of epicardial pacing leads for biventricular pacing therapy. In other words, patients who cannot achieve optimal resynchronization through transvenous lead placements will likely have accessible, exposed myocardium in ideal epicardial locations.

3.7 Conclusion Here, we provide one of the most complete characterizations of cardiac epicardial adipose distributions and thicknesses. Epicardial adipose deposits on these human hearts presented with characteristic patterns: i.e., with thick accumulations along the interventricular septa and the right ventricular margins, and thin accumulations on the diaphragmatic surfaces, left ventricular free walls, and right ventricular outflow tracts. These results provide a more complete anatomic understanding for those performing electrophysiologic procedures within the pericardial space.

Chapter 4: Electrical Parameters of Physiologic

His-Purkinje Pacing Vary by Implant Location in a

Canine Model

Alexander R. Mattson, BS1,2,3, Elizabeth Mattson, BS1,2,3, Mary-Lauren Mesich,

DVM, DACVS-SA2, Zhongping Yang2, PhD, Paul A. Iaizzo, PhD1,3

1Department of Biomedical Engineering, University of Minnesota, Minneapolis,

MN USA

2Medtronic, Mounds View, MN USA

3Department of Surgery, University of Minnesota, Minneapolis, MN USA

4.1 Preface The bundle of His serves as the conduit through which electrical conduction spreads from the atrial chambers past the central fibrous body and into the ventricles. Pacing the bundle of His results in a more physiologic conduction pattern, as the electrical impulse utilizes the native conduction system.

Many clinicians believe that His bundle pacing is the way of the future, as several studies have shown significant patient benefits in small, non-randomized trials.

However, there are hurdles to developing pacemaker technologies that can target a relatively limited area. In this study, we look at the anatomic dependence of His bundle pacing. We correlate implant location with electrical characteristics of the resultant paced beats, to further clarify the optimal targets for physiologic His bundle pacing.

Elizabeth Mattson contributed to study design, data analysis, and manuscript editing preparation. Zhongping Yang contributed to study design and manuscript editing and preparation. Paul Iaizzo contributed to study design and manuscript editing and preparation. Mary-Lauren Mesich contributed histologic analysis and manuscript editing and preparation.

This chapter was submitted for publication as an original manuscript in

HeartRhythm.

4.2 Synopsis

Background

Permanent His bundle pacing (HBP) is an attractive, perhaps more physiologic, alternative to traditional right ventricular pacing.

Objective

In this study, we utilized direct visualization to more comprehensively understand the anatomy central to HBP, correlating electrical lead performance to implant locations along the His bundle (HB) pathway.

Methods

Canine hearts (n=5) were isolated and reanimated using Visible Heart® methodologies. Medtronic 3830 SelectSecure™ leads were fixated where His potentials were present; the location of each implant was mapped/binned into four regions approximately analogous to the proximal, penetrating, and distal HB.

Locational differences in HBP capture and resultant QRS morphology were assessed.

Results

Average HBP capture thresholds did not significantly vary with respect to implant location (p=0.48, 1.0 ms pulse width). The resulting QRS morphologies from HB- paced beats varied in relation to implant location. As leads were placed further distally along the HB, the ratio of paced to native QRS complex duration increased (∆QRSpaced/∆QRSnative ratios were: Region 2, 0.84 ± 0.16; Region 3,

1.04 ± 0.42; Region 4, 1.74 ± 0.86). 90

Conclusion

We demonstrated correlation between the anatomical locations of HBP lead placement and resultant QRS morphologies in a reanimated canine heart model.

Proximal placement along the HB pathway resulted in more favorable QRS morphologies, suggesting improved selective HBP capture, with no significant increase in HBP capture thresholds. Pacing the HB in more proximal pathway locations improved the selectivity of HBP, and may confer electrical and anatomic benefits relative to distal HBP.

Keywords

His bundle; His-Purkinje; membranous septum; direct visualization

4.3 Introduction

Direct His bundle pacing is an attractive, physiologic alternative to traditional right ventricular pacing. Physiologic pacing requires the precise placement of a pacing lead onto the bundle of His, a branch of conductive fibers extending from the distal AV node, through the central fibrous body before branching into the Purkinje network in the high ventricular septum. The anatomy of the bundle of His varies between patients, potentially yielding different electrophysiological pacing profiles.84,85 The exact anatomic placement of His- pacing leads from patient to patient is often unclear. There have even been debates surrounding the location of His bundle pacing leads with respect to the tricuspid annulus, with previous studies analyzing lead placements during autopsy and with clinical imaging.26,86 Here, we use direct visualization to provide a comprehensive understanding of the anatomy central to His bundle lead implantation.

Permanent His bundle pacing was first described in the year 2000, in a subset of patients with chronic atrial fibrillation and dilated cardiomyopathy.21

Since then, physiologic pacing has become a popular topic of study in cardiac electrophysiology circles. When compared to traditional left ventricular leads, His bundle pacing has been hypothesized to improve left ventricular functional indices, normalize left ventricular dimensions, and lower NYHA functional class.27,28 Additionally, direct His bundle pacing is associated with both acute improvement in functional hemodynamics and long-term reduction of pacing- 92

induced cardiomyopathy and heart failure hospitalization rates when compared to right ventricular apical pacing.26,87

While pacing the bundle of His may provide acute electrical and functional improvements over traditional right ventricular apical pacing, is all permanent His bundle pacing created equally? Anatomic and physiologic intuition suggests that there may be differences in His bundle pacing when capturing the AV nodal transition zone versus the proximal bundle, penetrating bundle, distal bundle, and/or left and right bundle branches.88 This study looks at the effect of implant location along the His bundle on acute electrical function in reanimated canine hearts. We define correlations of acute electrical characteristics to pacing lead implant site, using direct visualization to place multiple His bundle pacing leads in each heart.

4.4 Methods Canine hearts (n=5) were isolated and reanimated using previously described Visible Heart® methodologies.89 Electrodes were placed on a supportive sponge to provide ECG vectors analogous to the traditional limb leads

(i.e. lead I,II,III). Additionally, temporary pacing electrodes (Medtronic 6495 bipolar temporary myocardial pacing electrode, Medtronic, Minneapolis, MN) were placed along the lateral wall of the right atrium, and the right ventricular outflow tract to provide local electrical signals from atrial and ventricular myocardium. Medtronic 3830 SelectSecure™ leads were fixated along the atrial

septum where HB potentials were present; the location of each placement was binned into one of four categories defined by proximity to the membranous septum (Figure 4.1). As such, leads were fixated in regions approximately analogous to the proximal, penetrating, and distal bundle of His. ECG, electrogram, and pacing capture thresholds (PCT) were collected at each implant site. Locational differences in electrical parameters were statistically assessed using a single factor analysis of variance with Tukey’s honest significant difference post-hoc comparison.

After all leads were placed, each heart was perfusion-fixed in 10% formalin solution using previously described methodologies. Histological

sections were made at each implant site; each section stained with Masson’s

Trichrome stain to test for the presence of conduction system components.

Notably, distinct helical tracts from implant were not visualized in these acute, ex- vivo implantations, thus depth of implant with respect to the cardiac conduction system was not assessed.

4.5 Results Pacing Capture Threshold Electrical signatures were characterized with respect to implant location. Average pacing capture threshold varied minimally with respect to implant location.

Figure 4.2 shows the average strength-duration curves for each implant region

(n=4 leads region 2, n=3 leads region 3, n=3 leads region 4). Notably, no implants within region 1 captured the His-Purkinje system.

Resultant QRS Morphology

** **

Region 2 Region 3 Region 4

The resulting QRS morphology from HB-paced beats varied as a function of implant location. Paced beats were analyzed in relation to native sinus beats by taking the ratio of QRS widths in the paced signals (∆QRSpaced) to native QRS widths (∆QRSnative). In region 1, despite a single implant showing HB EGM signature, no implant provided pure Hisian or para-hisian pacing capture. As leads were placed further distally, ∆QRSpaced increased in relation to ∆QRSnative

(∆QRSpaced/∆QRSnative ratios were: Region 2—0.84 ± 0.16; Region 3—1.04 ±

0.42; Region 4—1.74 ± 0.86). Further, 4/5 lead implants within region 2 showed characteristics of selective His bundle capture, while ¼ of implants in region 3 and 2/4 implants in region 4 were characterized as having selective capture.

Electrical Signal Timing

Figure 4.6: A—The time interval between the His potential and the takeoff of the QRS complex in the native electrogram decreases moving distally along the bundle of His. In region 1, His to QRS interval was the largest, averaging 49ms. In region 4, below the tricuspid annulus, His to QRS timing averaged 35ms. B—The time interval between pacing spike and QRS takeoff decreases moving distally along the bundle of His. Here, this time interval is defined as the time between pacing spike and the end of the delta wave, so to isolate Purkinje conduction. As such, leads implanted in region 4, below the tricuspid annulus, show a shorter conduction time between pacing spike and ventricular contraction. Conversely, leads implanted in regions 1,2, and 3 above the tricuspid annulus show longer intervals from pacing impulse to QRS initiation.

Histologic Findings Histological sectioning of each canine heart characterized the nearest component of the cardiac conduction system in each of the four implant regions.

In region 1, the proximal AV node was the nearest conduction-system component present, with nodal fibers present in approximately 50% of sections in region 1. In region 2, the nearest conduction system component was either the distal compact AV node or proximal bundle of His. Region 3 implants were most commonly positioned closest to the distal bundle of His, while region 4 implants were closest to the right bundle branch, just distal to the branching of the main

bundle of His. An example histologic sectioning with juxtaposed macroscopic lead locations is provided in Figure 4.7.

100

proximity to the implant site of Lead 1. Image C & G: The atrioventricular node (AVN) is nearest to the implant site of Lead 2. Image D & H: The His bundle (HB) is nearest to the implant site of Lead 3. Image E & I: The right bundle branch (RBB) at the His bundle bifurcation is nearest to the implant site of Lead 4. Ao=Aorta, AVN=atrioventricular node, CS=coronary sinus, IVC=inferior vena cava, HB=His bundle, LA=left atrium, LBB=left bundle branch, LV=left ventricle, RA=right atrium, RAA=right atrial appendage, RV=right ventricle, TV=tricuspid valve

4.6 Discussion Here, we demonstrate a correlation between implant locations along the bundle of His and pacing performance of the implant. In this series of animals, direct visualization aided in accurate placement of multiple leads along the length of the bundle of His, as well as precise anatomical characterizations of each implant. By binning implant locations into four major categories, based on anatomical region, we were able to approximate the results of pacing therapy at multiple loci along the bundle of His. We were able to achieve direct His bundle pacing capture along a considerable stretch of myocardium (several centimeters) at sites on both the atrial and ventricular side of the tricuspid annulus. For this study, region 1 was defined in the posterior segment of the Triangle of Koch, meaning that implants in this region are more likely to be placed in the slow conducting fibers of the AV node than the bundle of His. Indeed, only one lead placement across five animals showed distinct His bundle signature. In general,

His potentials were not mapped, and His bundle capture was not achieved in region 1. Further, histologic sectioning did not find evidence of His-Purkinje fibers in region 1.

101

As such, regions 2-4 are the most likely regions in which His bundle capture may be achieved. Notably, regional definitions are based on macroscopic definitions of each location with respect to the coronary sinus, tendon of Todaro, and antero-septal commissure of the tricuspid valve. Thus, rigorously defining a lead as being placed in the proximal or penetrating bundle of His upon implant is challenging, as regional anatomy varies from animal-to- animal. Instead, it is more accurate to define electrical characteristics of each region by saying that region 2 is more proximal than region 3, which is more proximal than region 4. In this study, histologic staining of the implant region guided mapping between macroscopic regional definitions, and the functional distinction of proximal vs penetrating vs distal bundle of His. Here, histologic staining confirmed that implant region 2 most commonly mapped to the proximal bundle of His or distal AV nodal fibers, region 3 mapped to the main body of the bundle of His, and region 4 mapped to more distal His bundle fibers often past the branching point into the left and right bundle branches.

Electrogram timing from the implanted leads gives additional information regarding implant location along the bundle of His. The timing intervals between

His potential and/or pacing impulse and the upstroke of the QRS complex were found to decrease as leads were implanted further distally. Intuitively, this increased delay corresponds to the time that a pacing impulse or native electrical signal takes to conduct along the length of the bundle of His. In a canine model, the unbranched portion of the bundle of His measures between 22 and 29 mm in

102

length.90 Conduction velocities in the bundle of His have been measured between 1.3-1.7 m/s.91 As such, we expect the timing delay between proximal and distal His bundle electrograms to be between 13-22ms. Here, we found a

~50ms delay between the native His potential in region 2 and QRS, and ~35ms delay between the native His potential in region 4 and the QRS. The 15ms time interval between these measurements corresponds well with the predicted time course of electrical conduction through the bundle of His. Thus, implants in regions 2 and 4 defined here likely correspond to the proximal and distal bundle of His. Histology confirms these findings. Histologic sectioning within region 2 shows the presence of the proximal bundle of his and/or distal atrioventricular node within the transition region. Sections from region 4 show implant locations just distal to the branching of the main bundle of His (Figure 4.8).

103

into the right and left bundle branches. HB = bundle of His, LBB = Left Bundle Branch, RBB = Right bundle branch.

In this study, implanted leads displayed physiologic evoked QRS complexes when implanted proximally along the bundle of His. This was evidenced by an increase in the ratio of paced to native QRS seen moving from region 2 to region 4 (Figure 4.3). Multiple factors may influence this finding. It is possible that pacing at more proximal loci along the His bundle results in a more complete recruitment of the entire His-Purkinje system leading to a more physiologic evoked response. For example, leads implanted within region 2 may be able to capture proximal His bundle fibers, thus activating both right and left bundle branches and the entirety of the subsequent Purkinje network (Figure

4.9A). However, leads implanted below the valve annulus in region 4 may only innervate the right bundle branch, and a portion of the left bundle branch tree, thus leading to less intrinsically physiologic conduction (Figure 4.9B).

These relationships can be seen in the animal case study shown in Figure

4.5. In this example, a lead implanted in region 2 is placed directly on the HB, proximal to the AV nodal junction. Here, selective HBP capture is achieved, and the native QRS complex is reestablished post-pacing. However, a lead implanted in region 4 shows a widened non-selective HBP capture morphology.

Histology and electrical recordings for this lead placement were consistent with primary right bundle branch capture, followed by retrograde left fascicular activation; i.e. the timing from pacing spike to QRS end was greater for this

104

implant placement than the timing between the native His bundle EGM and QRS end.

Additionally, implants in regions 2 and 3 may show reduced QRS complex width because of relatively little capture of ventricular myocardium.

Regions 2 and 3 in this study sit above the tricuspid valve annulus; thus, any pacing impulse from a lead implanted in these regions would need to travel through the insulating central fibrous body to capture ventricular myocardium. Of note, the relationship between implant location along the bundle of His and resultant QRS width is dependent on the degree and location of fascicular block

(Figure 4.9C&D); this experiment was performed in a healthy animal model, thus the conclusions derived pertain to healthy, native conduction system anatomies.

Importantly, the presented results likely do not translate to every patient anatomy. For example, in the setting of fascicular block of the left bundle branch, pacing the proximal bundle of His may confer less benefit than distal pacing

(Figure 4.9).

The noted changes in QRS morphology were uncoupled to pacing capture threshold in this experiment. Pacing capture threshold was not dependent on implant location, with His bundle capture being achieved at 3.7±1.8V, 2.3±0.6V,

3.2±0.7V in regions 2, 3, and 4 respectively. Notably, while mean pacing capture threshold did not vary significantly, proximal His bundle pacing was associated with a greater variance in pacing capture threshold. In region 2, this variance

105

likely results from lead implants commonly being placed near the transition zone of the atrioventricular node into the bundle of His. When placed too proximally within region 2, a higher capture threshold is needed to extend the pacing electric field beyond the distal AV node and into the proximal bundle of His.

A B

C D

106

There are notable differences between the canine and human His bundle anatomies. The canine bundle of His is significantly longer than that of the human.92 In addition, the bundle of His in the human heart is located adjacent to the antero-septal tricuspid commissure, slightly more anterior than that of the canine.92 As such, the length of paceable His bundle substrate along the proximal bundle of the human model may be less than that of the canine.

Nonetheless, the overall electrical and physiologic findings presented here should hold true across models.

Electrical performance favors lead placement above the tricuspid annulus

(in regions 2 and/or 3). Notably, lead implants in regions 2 and 3 also confer an anatomic advantage by not crossing the tricuspid annulus. It is conceivable that leads placed in region 4, or elsewhere near the distal bundle of His may impinge or penetrate the septal leaflet of the tricuspid valve. Acutely, this impingement could lead to increased tricuspid valve regurgitation and related clinical sequalae.

Conversely, a lead placed through valve tissue in region 2 shows no acute effect on tricuspid leaflet coaptation. In a chronic setting, entanglement in the tricuspid valve apparatus can augment valve insufficiency, or even prevent extraction of the lead. As such, the evidence here supplements the basic anatomic intuition that His bundle pacing achieved above the tricuspid annulus is superior to that below the annulus.

107

His bundle pacing implants placed in the right atrium may create challenges in programming appropriate sensitivity settings, due to the relatively large ratio of atrial to ventricular electrogram amplitude. In cases where the atrial electrogram exceeds the amplitude of the ventricular electrogram, atrial oversensing may result. Here, in reanimated canine hearts, the ratio of far field atrial electrogram amplitude to ventricular electrogram amplitude was generally small when leads were implanted on the ventricular side of the tricuspid valve (Figure 4.5, Region

4). On the atrial side of the valve, the ratio increased to varying degrees, though in most cases, the ventricular electrogram amplitude exceeded the atrial electrogram amplitude (Figure 4.5, Regions 2 and 3 serve as counter-examples).

In all, the potential benefits of proximal His bundle pacing must be weighed against the potential downsides on a case-by-case basis.

As previously discussed, supra-annular His bundle real estate in the human is likely reduced from what is reported here. Indeed, the detailed anatomic regions 1-4 may not correlate to clinical locations where His bundle pacing capture is achievable. However, overall, the presented findings indicate a potential benefit of proximal His bundle pacing as compared to distal, in appropriate patient demographics (one such counter example provided in 4.9).

In a clinical setting, this may potentially be achieved by implanting the His bundle pacing lead in a region where the timing interval between the His and ventricular electrograms is maximized.

108

4.7 Conclusion We demonstrated correlation between the anatomical locations of His bundle pacing lead placements, and resultant QRS morphologies in a canine model. Proximal placements along the bundle of His were associated with favorable QRS morphologies, suggesting more selective His bundle capture, with no significant increase in pacing capture threshold. Pacing the bundle of His in more proximal locations improves the selectivity of His bundle capture and may confer electrical and anatomic benefits to distal His bundle pacing.

109

Chapter 5: Visible Heart Visualization of

Physiologic His-Bundle Pacing and Surrounding

Anatomy within Reanimated Human Hearts

Alexander R. Mattson, BS1,2, Zhongping Yang, PhD2, Paul A. Iaizzo, PhD,

FHRS1

1Department of Biomedical Engineering and Surgery, University of Minnesota,

Minneapolis, MN USA

2Medtronic, Mounds View, MN USA

110

5.1 Preface The previous chapter provided a link between the anatomic location of His bundle pacing lead implant and the electrical characteristics of the resulting paced beats. Here, we begin to bridge the gap between animal models and clinical implants. In this study, we attempt to recreate native conduction patterns by pacing at distinct loci along the bundle of His in the human heart. We describe that the success of a His bundle pacing lead is intricately tied to each individual’s anatomy.

Zhongping Yang and Paul Iaizzo contributed to experimental conception, design, and implementation as well as manuscript editing and preparation.

5.2 Synopsis Introduction Permanent His bundle pacing (HBP) is an attractive, perhaps more physiologic alternative to right ventricular pacing. The anatomy of the His bundle (HB) varies between patients, potentially yielding different electrophysiological pacing profiles. Here, we use direct visualization in reanimated human hearts, to provide a comprehensive understanding of the anatomy central to HBP lead implantation.

Case Report Two human donor hearts (53F, 50F, neither with conduction abnormalities or significant arrhythmogenicity) were reanimated and imaged using previously described Visible Heart Methodologies. Donor hearts were instrumented with surface ECG leads and temporary bipolar pacing leads to record atrial and

111

ventricular electrograms. A C304 steerable sheath was used to deliver

Medtronic 3830 Select Secure leads to the lower border of the membranous septum. His bundle potentials were mapped, and leads were implanted in regions where selective His bundle pacing capture was achieved.

Discussion and Conclusions Here, we directly image the link between anatomic placement of a HBP lead, and the electrophysiological output. We show that pacing within the right atrium along the proximal lower border of the membranous septum was able to induce selective His bundle capture in both donor hearts. Further, we show that in only one donor heart, selective His bundle pacing capture was achieved by leads placed both above and below the tricuspid annulus. As such, lead implant location along the membranous septum plays an important role in the selectivity of His bundle pacing in the human heart. Although lead placement using direct visualization is not representative of the visualization techniques available in a clinical setting, the images presented here have notable educational value for both clinicians and design engineers.

112

5.3 Introduction

Permanent His bundle pacing is an attractive, perhaps more physiologic alternative to right ventricular pacing. Physiologic pacing requires the precise placement of a pacing lead onto the bundle of His (HB), a branch of conductive fibers which extends from the distal AV node, running along the membranous septum, before diving into the ventricular septum. The anatomy of the HB varies between patients, potentially yielding different electrophysiological pacing profiles. Given that HBP capture may be achieved from either the atrial or ventricular side of the tricuspid annulus, there is discussion surrounding the benefits and downsides of each. Here, we use direct visualization in reanimated human hearts, to provide a comprehensive understanding of the anatomy central to HBP lead implantation.

113

5.4 Case Report We present the reanimations of two non-viable human donor hearts, whose donation for research was supported via our local organ procurement agency (LifeSource, Minneapolis, MN). Information regarding donated specimens, including video footage of specific anatomic features may be found at the University of Minnesota’s Atlas of Human Cardiac Anatomy free access website, vhlab.umn.edu/atlas/, referenced by patient number. Notably, in both cases presented here, donors presented with intact AV conduction and non- diseased cardiac electrophysiology.

The donated hearts were reanimated using previously detailed Visible

Heart® methodologies.89 Endoscopic cameras (IplexFX, Olympus Corporation,

Tokyo, Japan) placed within the heart allowed direct anatomic visualization during the implant.

Each heart was instrumented to give a surface ECG analog. Of note, ECG recordings in this study are taken from leads placed directly adjacent to the heart, achieving electrical continuity through a conductive gel. While analogous to clinical recordings, the electrical vectors shown in this case are not precisely representative of a clinical ECG. Temporary pacing leads implanted in the right atrial appendage and right ventricular outflow tract provided local RA and RV electrograms.

114

A videoscopic landscape of the tricuspid annulus between the coronary sinus and antero-septal tricuspid commissure was taken, so to identify the approximate location of the HB. A 6mm endoscope placed into the root of the ascending aorta on the right coronary cusp, illuminated the membranous septum within the right atrium (Figure 5.1A).

In each case, the HB was mapped using recordings from a 3830 lead

(Medtronic, Minneapolis, MN, USA) through a C304 steerable sheath.

Video recordings of both implant procedures may be found in the supplementary material (Supplementary Video 5—His Bundle Pacing in

Reanimated Human Hearts).

Patient 462

Patient 462 was a 53 year old female, with a history of hypertension and mild calcification in the left anterior descending coronary artery. In this patient,

HB potentials were mapped and HBP capture was achieved in only one location.

The 3830 lead was implanted just inferior to the most proximal segment of the membranous septum, on the atrial side of the tricuspid valve (Figure 5.1A).

Native electrograms from the implanted lead placement showed a distinct His signature. Selective HBP capture was achieved at low capture threshold (Figure

5.1C): 1.7V at 1.0ms, 2.3V at 0.5ms, and 3.8V at 0.2ms. When pacing at 1.0 ms pulse width, fusion capture was seen at voltage greater than 7V. QRS width was not significantly different between native sinus and selective HBP beats (103ms native, 101ms HBP).

115

A CS MS

TV Septal Leaflet

B C

Figure 5.1: Endoscopic footage within the right atrium of patient 462. A Medtronic 3830 lead (Medtronic, plc Minneapolis, MN) was fixated just below the lower border of the membranous septum. (B) ECG Leads I, II, AEGM, VEGM, Bipolar 3830 recording for native sinus beats. Atrial and Ventricular EGMs were taken from temporary pacing leads placed in the right atrial appendage and the right ventricular outflow tract. CS—Coronary Sinus; TV—Tricuspid Valve; MS—Membranous Septum

116

Patient 475

Patient 475 was a 50 year old female, with no salient cardiac history. In this patient, HBP capture was achieved in two separate locations, spanning the atrial and ventricular side of the tricuspid valve. The first lead was implanted inferior to the proximal membranous septum on the atrial side of the tricuspid annulus (Figure 5.2A). Here, stable, selective HBP capture was seen (Figure

5.2E); thresholds were 1.9V at a 1.0ms pulse width, 2.8V at 0.5ms pulse width, and 4.2V at 0.2ms pulse width. Fusion capture was seen at outputs higher than

6V at a 1.0ms pulse width.

A second HBP lead was placed on the ventricular side of the tricuspid annulus, just below the membranous septum. In this location, selective HBP capture was achieved at low output: 2.0V at a 1.0ms pulse width, 2.3V at a

0.5ms pulse width, and 6.3V at a 0.2ms pulse width. Again, fusion capture was seen at higher output.

117

A B

C D

E F

118

5.5 Discussion and Conclusions

Here, we directly image the link between anatomic placement of a HBP lead, and the electrophysiological output. In both of the presented cases, an atrial lead position, just inferior to the most proximal visible segment of the membranous septum achieved selective HBP capture. However, were we able to successfully map and pace the HB on the ventricular side of the tricuspid annulus in only one of two cases. In both patients, selective HBP capture was achievable at low output, while fusion capture was seen at high outputs, suggesting lead placement within His-Purkinje tissues.

Macroscopic anatomy of the HB has been previously detailed by

Kawashima and Sasaki.84 In their study, HB anatomies are binned three distinct anatomical categories:

1. Type I (47% of hearts): The HB is surrounded by a thin layer of myocardial

tissue. The HB courses along the lower border of the membranous

septum.

2. Type II (32% of hearts): The HB is insulated by a thicker layer of

myocardial fibers. The HB runs separate from the lower border of the

membranous septum, with a discrete separation between the two.

3. Type III (21% of hearts): The HB has no surrounding myocardial fibers

(bare) and runs just beneath the endocardial surface.

119

Some have hypothesized that the anatomic variation described results in the variable selectivity of HB capture seen in the clinic. Possibly the most common clinical HBP scenario involves non-selective HBP capture with ventricular fusion at high output pacing, with selective HBP occurring as output is lowered. This electrophysiologic phenomenon corresponds best to a Type I anatomy, where the lead is placed directly within His-Purkinje tissue. In this case peripheral activation of myocardial fibers would be seen at high voltage, with selective HBP capture with low voltage.

In other clinical cases, non-selective fusion capture is achieved regardless of voltage—a potential result of Type II anatomy. Here, the thick myocardial fibers may prevent selectively targeting the HB. Type III anatomy, where the HB is readily accessible, just beneath the myocardium, may drive cases where selective his capture is achieved independent of pacing voltage output.

In these two cases, the electrophysiological response mimics the hypothesis for a Type I anatomy; i.e. selective His capture at low output, fusion at high output. This is consistent with the macroscopic anatomical imaging of the lead placements. Here, we show the leads fixated directly to the lower border of the membranous septum, in both cases. In a Type II anatomy, we would expect a larger separation between the lower border of the membranous septum and the lead helix, as well as a different electrophysiologic capture profile. In patient 462, we were able to map his signals in only a small region above the tricuspid

120

annulus, and not at all below the valve plane. With a naked HB, as seen in Type

III anatomy, we would anticipate a larger region along the lower atrial septal border to show native His signals. The selective HBP capture seen beneath the tricuspid annulus in patient 475 suggests that the patient’s anatomy is relatively more similar to Type III anatomy, compared to patient 462.

In these cases, direct visualization aided in the positioning of the leads; indeed, live video was the primary imaging modality used in lead placement.

Although lead placement using direct visualization is not representative of the visualization techniques available in a clinical setting, the images presented here have notable educational value for both clinicians and design engineers.

121

Chapter 6: 3-Dimensional Anatomic Assessment of the Human Right Atrial Appendage: Implications for

Atrial Fixation Mechanisms

Alexander R. Mattson1,2, BS

1University of Minnesota, Departments of Biomedical Engineering and Surgery

2Institute for Engineering in Medicine, University of Minnesota

122

6.1 Preface Current leadless technologies are approved only for right ventricular—

VVI(R)—pacing. To become a truly transformative technology, leadless pacemakers must be adapted for the diverse indications of bradycardia pacing.

If next-generation leadless pacemakers are to be placed in the right atrium, the right atrial appendage is a likely target; the appendage offers relatively straight-forward access from the inferior vena cava and serves as the primary location for current atrial lead placements.

Designing a leadless device for the right atrium necessitates a comprehensive understanding of the atrial anatomy. This chapter provides key anatomic detail of the right atrial appendage: a potential anatomic target for next- generation leadless pacing.

6.2 Synopsis Background Any fixation mechanism within the right atrium which hopes to anchor to pectinate must consider the wide spacing between the pectinate muscles of the human heart. Leadless pacing fixation mechanisms must be able to consistently span the gaps of pectinate muscle to reliably fix within the right atrial appendage,

123

without engaging the thin-walled visceral pericardium. Here, we aim to define the thickness and spacing of the pectinate muscle within the human heart.

Methods Twenty-one human hearts were MRI scanned in high resolution (0.1mm slice spacing). 3-Dimensional reconstructions of the atrial anatomy were created using

Mimics and 3-Matic software (Materialise, Leuven, Belgium). Appendage sizing and pectinate spacing were quantified within Mimics software.

Results Within this sample of hearts, pectinated regions covered 53.0±5.0% of the right atrial surface. Appendage size increased with patient age, height, and weight.

Pectinate muscles become thinner when moving anterior and medially from the lateral wall of the right atrial appendage (3.48±1.36 mm, 2.63±1.16 mm,

1.73±0.76 mm for lateral wall, appendage roof and appendage tip, respectively).

The pectinate network becomes denser moving anteriorly from the lateral wall

(5.02±1.07mm, 4.44±1.61mm, 2.60±0.70mm average spacing between major pectinate muscles for lateral wall, appendage roof, and appendage tip, respectively).

Discussion and Conclusion Three-dimensional right atrial appendage anatomy plays a critical role in the successful implementation of any atrial fixation mechanism. This analysis quantitatively defines the amount of pectinated surface area and spacing between major pectinate muscles within the human heart. 124

6.3 Background Anatomy Introduction The right atrium of large mammalian species

contains an appendage (auricle) that is believed to

be a vestigial remnant of fetal development. In the

human heart, the right atrial appendage is located

anterior and medial of the central atrial body,

beginning proximal to the systemic venous sinus

(i.e. the inferior vena cava and superior vena cava

ostia and coronary sinus ostium), wrapping around

Figure 6.1: The right atrial the lateral and anterior portions of the right atrium, appendage of the human heart. and overlapping the root of the aorta at its most distal tip. Within the body of the right atrium, the junction of the right atrial appendage to the smooth-walled vestibule of the right atrium is demarked by the crista terminalis and the vestibule of the tricuspid valve. The right atrial appendage is lined with pectinate muscles, which web across the body of the appendage, oriented primarily perpendicular to the vestibule of the tricuspid valve. The pectinate muscles pass across the thin visceral pericardium separating the appendage from the pericardial space. Notably, the epicardial wall in the right atrial appendage is so thin that a business card placed on the epicardial surface can be read when looking from inside the right atrial appendage (Figure 6.2).

125

Figure 6.2: (right) A business card is read through the body of the right atrial appendage. Thicker pectinate muscle bands span the clear visceral pericardial layer.

Central Anatomic Questions Any fixation mechanism within the right atrial appendage which hopes to anchor to pectinate must consider the wide spacing between the pectinate muscles of the human heart (see Figure 6.2). Leadless pacing fixation mechanisms must be able to consistently span the gaps of pectinate muscle to reliably fix within the right atrial appendage, without engaging the thin-walled visceral pericardium,.

Here, we aim to define the thickness and spacing of the pectinate muscle within the human heart.

6.4 Methods

Acquisition and Imaging of Human Heart

Specimens

Twenty-one human hearts were donated to the

University of Minnesota’s Visible Heart

Laboratory from a local organ procurement Figure 6.3: High resolution MRI slice from a human heart at the University agency (Lifesource, Minneapolis, MN). of Minnesota's Visible Heart Lab

126

Donated hearts were perfusion-fixed in a 10% formalin solution, preserving the atrial anatomy in an approximation of atrial diastole. Fixed hearts were gelled in a 7% agar gel to prevent motion artifact during the acquisition of a T1-weighted

MRI scan. As the specimens were isolated and stationary, MRI scans returned imaging with anatomic resolution of 0.1 mm. An example MRI slice from these high-quality scans can be seen in Figure 6.3.

3-Dimensional reconstructions of the atrial anatomy were created using Mimics and 3-Matic software (Materialise, Leuven, Belgium).

Surface Area of Pectinated Region of RAA

127

Example 3-Dimensional reconstructions of the human atrial anatomy can be seen in Figure 6.4 and Figure 6.5. In the reconstructions, the 3-matic Mark tool was used to highlight and quantify surfaces within the right atrium (Figure 6.4 and

Figure 6.5). Both pectinated and non-pectinated regions within the right atrium were defined. When identifying the pectinated regions within the atrium, the epicardial surface was highlighted to prevent artificially increasing the pectinated surface area via highlighting pectinate mounds (i.e. the “bumpiness” of the pectinated region leads to artificial increases in measured surface area), see

Figure 6.5.

The total surface area of the right atrial appendage (i.e. all pectinated regions) was defined and characterized as a percentage of total atrial surface area.

128

Pectinate Size and Spacing

Pectinate muscle size and spacing was measured using the measure and ellipse tools within the Mimics anatomic modeling software package. In each specimen

MRI scan, 3 axial slices were taken, evenly spaced across the appendage. For each slice, three representative pectinate muscles were chosen: one muscle on the lateral wall, one on the appendage roof, and one at the appendage tip. As such, in each specimen, nine total pectinate muscles were analyzed: three from the lateral wall of the appendage, three from the roof of the appendage, and three from the tip of the appendage.

Each pectinate muscle was measured for thickness, and proximity to adjacent pectinate muscles. Thickness was evaluated along the long and short axes of each pectinate muscle (Left Figure 6.6). Proximity was evaluated by measuring between the center of the pectinate muscle, and the center of its closest neighbors on either side (i.e. moving toward or away from the appendage tip, see left Figure 6.6). Measurements taken on the 2-Dimensional MRI slice were viewed in the 3-Dimensional reconstructions, to confirm measurement relevance to 3-Dimensional anatomy (Right Figure 6.6).

129

6.5 Results Donor Heart Demographics

Characteristics of the 21 donor hearts are presented in Table 6.1. Average patient age was 54.7 years. Patients were evenly distributed across gender, height range, and weight range. Three donors had a history of congestive heart failure, with an additional three having varied disorders related to the cardiac conduction system.

130

Table 6.1: Donor Patient Demographics/Characteristics Characteristic Number of Patients Gender Male 10 (47.6%) Female 11 (52.4%) Age µ = 54.7 ± 17.9 years 0-30 years 4 (19.0%) 31-60 years 6 (28.6%) 61+ years 11 (52.4%) Height µ = 166.2 ± 14.9 cm < 160cm 4 (19.0%) 160cm-170cm 8 (38.1%) 170cm -180 cm 5 (23.8%) >180 cm 4 (19.0%) BMI µ = 27.6 ± 5.7 <25 (Underweight/Normal 8 (38.1%) Weight) 5 (23.8%) 25-30 (Overweight) 8 (38.1%) >30 (Obese) Disease State Hypertension 8 (38.1%) Congestive Heart Failure 3 (14.3%) Atrial Fibrillation 1 (4.8%) Sinus Bradycardia 1 (4.8%) Right Bundle Branch Block 1 (4.8%)

131

Surface Area of Right Atrial Appendage

Within this sample of hearts, pectinated regions covered 53.0±5.0% of the right atrial surface. The ratio of pectinated right atrial surface area to total right atrial surface area did not vary with patient gender, age, height, weight, or BMI. These relationships are depicted in Figure 6.8.

Across this sample of hearts, the median surface area of the pectinated right atrial appendage was 53.5±14.4 cm2. The total pectinated surface area shows an increasing trend with patient height, weight, and age (Figure 6.8).

Within the small sample of patients with a related history of conduction system abnormalities (i.e. atrial fibrillation, right bundle branch block, sinus bradycardia, sinus tachycardia) there was a statistically significant increase in the surface area of pectinated area within the right atrium at a p=0.05 level: 65.9cm2 vs 53.5cm2, p

= 0.039. There was no difference in the ratio of pectinated to total surface area in these hearts (51.6% vs 53.0%, p = 0.70).

132

Ratio of Pectinated Surface Area vs Height Pectinated Surface Area vs Height

Ratio of Pectinated Surface Area vs Age Pectinated Surface Area vs Age

Ratio of Pectinated Surface Area vs Weight Pectinated Surface Area vs Weight

133

Pectinate Dimensions across the Right Atrial Appendage Pectinate Spacing by Location Pectinate Thickness by Location

Pectinate muscle thickness varied across the length of the right atrial appendage.

Pectinate muscles become thinner when moving anterior and medially from the lateral wall of the right atrial appendage (3.48±1.36 mm, 2.63±1.16 mm,

1.73±0.76 mm for lateral wall, appendage roof and appendage tip, respectively).

Despite thinning pectinate when moving from lateral wall towards the appendage tip, the pectinate network becomes denser, resulting in a closer spacing between major pectinate muscles (5.02±1.07mm, 4.44±1.61mm, 2.60±0.70mm for lateral wall, appendage roof, and appendage tip, respectively).

134

6.6 Discussion

Surface Area of Right Atrial Appendage

The pectinated region of the right atrium extends from the lateral wall to the right atrial appendage tip, comprising over 50% of the paceable surface area within the right atrium. This area includes the most common placements for atrial pacemaker leads.

Right atrial appendage size—defined here by surface area—increases as the size of the patient—as defined by height and weight—increases. However, the ratio of pectinated surface area to total surface area does not increase, suggesting that the right atrial appendage will maintain size in proportion to the right atrium as the heart grows. Importantly, patient size will correlate with the amount of paceable substrate for any device implanted within the right atrial appendage.

The surface area of the right atrial appendage increases with associated conduction system diseases. The evidence presented here corresponds to previous literature; abnormal atrial hemodynamic loading, such as that experienced during supraventricular arrhythmia, leads to dilation of the right atrium.

Pectinate Dimensions across the Right Atrial Appendage

Pectinate muscle thickness and spacing varies across the right atrial appendage. Broadly, pectinate muscles are wider and further spaced near the 135

lateral wall, while thinner, but closely packed at the tip of the right atrial appendage. As such, a leadless pacemaker placed in the lateral wall or roof of the appendage is more likely to encounter clear serous pericardium than placements near the right atrial appendage tip. For tined fixation mechanisms, this indicates higher probability of failed fixation when moving laterally along the right atrial appendage.

Importantly, the spacing between pectinate muscles in the atrial appendage results in a low probability that a tined mechanism will be deployed in a pectinate gap (i.e. in a region where the tines have no pectinate substrate to affix to).

Average spacing between pectinate in this analysis was 5.02 mm on the lateral wall (note: the maximum spacing seen in this analysis was 9.37 mm).

6.7 Conclusions

Three-dimensional right atrial appendage anatomy plays a critical role in the successful implementation of any atrial fixation mechanism. This analysis quantitatively defines the amount of pectinated surface area within the human heart. Additionally, it quantifies the thickness and spacing of the pectinate muscle, two pivotal anatomic considerations when designing a fixation mechanism for the right atrial appendage.

136

Section III. Biomechanics Informing Fixation of Next- Generation Pacing Systems

As previously highlighted, next generation leadless pacing systems must be adapted to fit new anatomic challenges. In particular, the drive for dual chamber leadless pacing necessitates leadless pacemakers capable of pacing in both the atrium and the ventricle. In this section of my thesis, I tackle two key biomechanical challenges that will aid in the design and implementation of next- generation leadless pacing systems.

Fixation within the right atrium Any fixation mechanism designed for implant within the right atrium must be able to leverage the unique biomechanics of the chamber; i.e. to appropriately fixate tissue while avoiding perforation. In the first chapter of this section, I highlight biomechanical perforation properties of the right atrium, identifying key relationships that may prove pivotal to fixation design within the right atrium.

Fixation Verification in the Right Atrium

Once a device is deployed within the right atrium, a clinician must verify appropriate fixation. The second chapter within this section dives into key challenges with fixation verification in the right atrium, providing further biomechanical information for device designers.

137

Chapter 7: Perforation Properties of the Right Atrial

Appendage

Alexander R. Mattson, BS1,2, Vladimir Grubac, BS2, Michael D. Eggen, PhD2,3,

Paul A. Iaizzo, PhD1,3

1Department of Biomedical Engineering, University of Minnesota, Minneapolis,

MN USA

2Medtronic, Mounds View, MN USA

3Department of Surgery, University of Minnesota, Minneapolis, MN USA

138

7.1 Preface In the drive for next-generation leadless pacing, appropriate fixation is paramount. Notably, despite lower overall complication rates compared to transvenous leaded systems, the potential complications from a leadless pacemaker dislodgment and/or perforation may present more severe clinical sequelae when compared to the analogous transvenous lead complications. The relative success of leadless pacemakers may be at least partially traced back to appropriate design of their fixation mechanisms. Next generation dual-chamber leadless pacemakers must be able to fixate into atrial tissues, which critically relies on the biomechanical properties of atrial tissue. This chapter of my thesis begins a biomechanical characterization of the right atrium, focusing on the relative perforation properties needed to penetrate pectinate muscle and perforate through atrial epicardium.

Vladimir Grubac and Michael Eggen contributed to experimental conception, design, and data analysis. Paul Iaizzo contributed to chapter editing and preparation. Data within this chapter was originally published as an abstract at the 2016 American College of Cardiology Annual Meeting. Portions of this chapter were originally published as a technical brief in the Journal of Medical

Devices:

Mattson AR, Grubac V, Eggen MD, Iaizzo PA, ”Acute perforation

properties of the right atrial appendage.” Journal of Medical Devices,

Transactions of the ASME, 10(2) DOI: 10.1115/1.4033147 139

7.2 Synopsis

Introduction The drive for dual chamber leadless pacing systems necessitates the design of novel fixation mechanisms, or the leveraging of existing mechanisms into the novel, atrial space. We performed the following studies to better define the underlying relationships between perforation/penetration forces of the right atrial appendage (RAA) and the relative surface areas of the applied penetrating devices to help inform design criteria for right atrial fixation mechanisms.

Methods The right atrial appendage was dissected from hearts of swine (n=24) and human organ donors (n=10); a custom-made chamber secured these samples.

Cylindrical penetrators, attached to a digital force sensing machine, were aligned with pectinate muscles and advanced at 120mm/min in a quasistatic fashion until perforations of the endocardium, pectinate myocytes, and associated epicardial layers (the visceral pericardium) of the RAA had occurred. The perforation properties of both pectinate muscle and associated epicardium of the RAA were investigated.

Results In swine and human hearts, the forces required to perforate pectinate muscle were significantly lower than the forces required to perforate the epicardial/visceral pericardial layer (p ≤ 0.05). The relationship between the surface area of the penetrator and force of perforation was found to be linear for both pectinate muscle and epicardial tissue. 140

Conclusions

Any designed active fixation mechanisms should take these relationships into account; e.g., balancing the amount of force and displacement required to actively engage the pectinate muscle while avoiding perforation of the associated epicardial layer (visceral pericardium). The described biomechanics have significant value to device engineers who hope to leverage tissue properties for mechanical design.

141

7.3 Introduction and Background

The drive for next-generation dual chamber leadless pacing systems necessitates either the design of novel fixation mechanisms, or the leveraging of existing mechanisms into a novel space. Notably, next-generation atrial leadless pacing systems will be targeting atrial tissues, which are significantly thinner than ventricular myocardium, potentially augmenting perforation risk. Further, rates of atrial lead dislodgements exceed the ventricular dislodgement rate.93 As such, any leadless device must be able to leverage the unique tissues within the atrium to provide adequate fixation.

To date, little research has been reported relative to the biomechanical properties of the right atrial tissues. Therefore, we performed the following studies to better define the underlying relationships between perforation/penetration forces of the right atrial appendage (RAA) and the relative surface areas of the applied penetrating devices (i.e. a lead or fixation mechanism). In the future, such information may be utilized to optimize design for devices placed within the right atrium.

142

7.4 Methods

Testing methodology was modified from protocols previously used in the design of leadless pacemaker fixation mechanisms.35 The right atrial appendage was dissected from hearts of Yorkshire Cross swine (n=24) and from hearts of human organ donors (n=10); a custom-made chamber, depicted in Figure 7.1, secured these samples. An oxygenated Krebs-Henseleit buffer circulated within this chamber maintaining tissue viability, and tissue temperature at 37oC.

Cylindrical penetrators with diameters between 0.254 mm-12.70 mm were attached to a digital force sensing machine (Chatillon, Ametek, Berwyn, PA) to allow measurement of force as a function of displacement during each penetration. Penetrators were aligned with pectinate muscles and advanced at

120mm/min in a quasistatic fashion until perforations of the endocardium, pectinate myocytes, and associated epicardial layers (the visceral pericardium) of the RAA had occurred. The perforation properties of both pectinate muscle and

143

associated epicardium of the RAA were investigated (for swine, n= 500 total penetrations: for human, n=300 total penetrations). Additionally, the appendage or pectinated regions of the right atrium in each heart were sectioned into thirds—referenced here as lateral wall, appendage roof, and distal appendage tip: i.e., to further investigate perforation force variations within each anatomical region. Obtained data were imported into Matlab Analysis software. The observed peak pectinate and epicardial perforation forces were manually selected, with the criterion that perforation required a rapid 5% drop in force.

7.5 Results The relationship between perforation force and penetrator cross-sectional surface area (CSA: mm2) was found to be linear for both pectinate muscle, and

144

epicardial tissue. Furthermore, across species, we determined the relationship between the perforation force (Fp, N) and cross-sectional area of the penetrator

(CSA, mm2) for the pectinate muscles, to be:

푆푤푖푛푒: 퐹푝 = 0.99 ∗ 퐶푆퐴 + 0.27, 푅2 = 0.62, 퐻푢푚푎푛: 퐹푝 = 4.13∗ 퐶푆퐴 + 0.14,

푅2 = 0.35 Equation 1 For the epicardial layer, we found the relationship to be:

푆푤푖푛푒: 퐹푝 = 1.38 ∗ 퐶푆퐴 − 0.60, 푅2 = 0.44, 퐻푢푚푎푛: 퐹푝 = 5.26 ∗ 퐶푆퐴 + 0.89,

푅2 = 0.24 Equation 2

Figure 7.3: The relationship between penetrator surface area and force at tissue penetration. On average, epicardial (thin wall) tissues take a greater amount of force to penetrate than pectinate muscle. In swine, for each penetrator size, the forces required to perforate pectinate muscle were significantly lower than the forces required to perforate the epicardial/visceral pericardial layer (p ≤ 0.05). In addition, the slope of pectinate perforation force relationship was significantly less than the epicardial perforation 145

force (p ≤ 0.05), where the 95% confidence intervals for the slopes were 0.86

N/mm2 to 1.07 N/mm2 and 1.44 N/mm2 to 1.80 N/mm2, respectively; see Figure

7.3A.

For the swine model, linear regressions for pectinate perforation forces, as a function of penetrator cross-sectional areas were calculated for each of the three anatomical sections within the right atrial appendage.

2 퐿푎푡푒푟푎푙 푊푎푙푙: 퐹푝 = 1.39 ∗ 퐶푆퐴 + 0.27, 푅 = 0.65 Equation 3

2 퐴푝푝푒푛푑푎푔푒 푅표표푓: 퐹푝 = 0.98 ∗ 퐶푆퐴 + 0.30, 푅 = 0.54 Equation 4

퐴푝푝푒푛푑푎푔푒 푇푖푝: 퐹푝 = 0.73 ∗ 퐶푆퐴 + 0.26, 푅2 = 0.64 Equation 5

Regressions differed significantly (p<0.05) between the distal tips of the appendage, roof of the appendage, and the lateral wall. 95% confidence intervals for the slope of the regressions were: 0.63 N/mm2 to 0.82 N/mm2, 0.72

N/mm2 to 1.20 N/mm2, and 1.18 N/mm2 to 1.60 N/mm2 respectively: indicating significantly different perforation properties of myocardial bundles at distal tip vs lateral wall. Perforations of the epicardial/pericardial layer within various anatomical RAA locations all required similar forces. Derived regressions for the performed epicardial perforations did not differ significantly across the three anatomical locations.

There was no clear relationship between penetrator surface area and displacement at tissue penetration. Pectinate muscle was penetrated, on

146

average, at approximately 5mm regardless of penetrator diameter. Similarly, epicardial perforation occurred, on average at 10mm, regardless of penetrator diameter.

The linear relationship between penetrator surface area and force of penetration, suggests that perforation forces may be normalized by surface area to create an overall cumulative probability of perforation given an applied pressure (Force/CSA, given that the penetrator exists within the CSA range tested). Similarly, plots of perforation probability under a given penetrator displacement may be generated.

Figure 7.4 Cumulative probability distributions of perforating through a given human heart tissue layer as a function of displacement and pressure applied. Here, pectinate muscle penetration is represented in light blue, while visceral pericardial penetration is shown in darker blue. Pressure and displacement values are normalized to the maximum perforation displacement/pressure of the visceral pericardial layer. At a normalized pressure of 0.25, 90% of all pectinate muscles were penetrated, while only 50% of epicardial layers were perforated. Similarly, at normalized displacement of 0.30, 60% of all pectinate layers were penetrated, while no perforations were seen.

147

7.6 Discussion This study has developed a reproducible benchtop approach to assess the relative perforation properties within the right atrial appendage of either human or swine hearts, with respect to applied penetrator surface areas. The study has determined the forces and displacements required to puncture the pectinate muscle as well as through the adhered epicardial/pericardial layers in multiple anatomical regions within a given RAA. This type of biomechanical tissue property information can be then utilized for optimizing future designs and implementations of novel atrial fixation technologies.

Perforations of either the pectinate muscles or epicardial tissues relate linearly to applied penetrator cross-sectional areas; pectinate punctures required significantly lower forces for all cylinder sizes. The relationship between force and cross-sectional area of the penetrator was found to be linear. As such, a cumulative probability of perforation distribution was created by pooling penetration tests from different sized penetrators. Notably, there is a distinct penetration pressure gap between pectinate muscle and visceral pericardium, with each given percentile of pectinate perforation occurring at smaller pressure than the corresponding pressure of epicardial penetration. As such, a fixation mechanism may be able to leverage the difference in these two layers (i.e. by attaining a column strength capable of penetrating the pectinate muscle, which buckles prior to epicardial perforation forces).

148

Interestingly, there is also a significant gap in terms of displacement between the pectinate and epicardial perforation populations. At a penetrator displacement of 5mm, more than 60% of all pectinate muscles were successfully penetrated, while no epicardial perforations were seen. Indeed, 50% of epicardial tissues were perforated only after applying twice as much penetrator displacement (10mm). This signals the importance of optimizing the displacement of a fixation mechanism within the right atrial appendage.

Epicardial/visceral pericardial perforation properties were found to be independent of the relative anatomical location within the RAA: whereas pectinate muscle in the distal tips of the appendage took significantly less force to perforate. These identify biomechanical properties deserves particular attention with the design of higher surface area fixation mechanisms.

7.7 Conclusion

149

Chapter 8: Assessing the Relationship between

Right Atrial Stiffness and Chamber Pressure to

Quantitatively Define Myocardial Tensile Properties

Alexander R. Mattson1,2,3, Michael D. Eggen PhD2,3, Vladimir Grubac3, Paul A.

Iaizzo PhD1,2

1Department of Biomedical Engineering, University of Minnesota

2Department of Surgery, University of Minnesota

3Medtronic, plc

150

8.1 Preface Once a pacemaker is fixated into tissue, it is critical to ensure that it is properly fixated; i.e. to minimize the risk of dislodgement and subsequent loss of pacing therapy. Fixation checks in transvenous pacing leads may be as simple as quickly removing the lead stylet or “wiggling” the lead tip. With leadless pacemakers, adequate fixation is critically important, as device embolization may present more severe clinical sequelae when compared to transvenous lead dislodgement. In the case of current leadless pacing systems, fixation is verified by applying a tensile force to the proximal end of the device. During these “tug” tests, tensile force applied to the device, causes it to displace slightly. With the

Medtronic MicraTM, any fixation tines engaged in tissue will deflect during this tensile test, indicating engagement with tissue. For other leadless systems, force buildup as the device is pulled may be felt, indicating appropriate fixation.

As leadless pacing systems begin to make the transition into the atrium for dual chamber pacing, new challenges will define fixation testing. The right atrium is exceptionally compliant in comparison to ventricular tissue. As such, any tug test performed in the right atrium will result in a lower amount of tensile force to the user for a given displacement. This chapter defines a test methodology to test the stiffness of right atrial tissues, defining key biomechanic parameters central to the design and implementation of next-generation leadless pacing systems.

151

Michael Eggen, and Vladimir Grubac contributed to experimental design, and manuscript editing. Paul Iaizzo contributed to manuscript editing and preparation.

The following chapter was published in the ASME Journal of Medical

Devices.:

Mattson AR, Eggen MD, Grubac V, Iaizzo PA. “Assessing the Relationship

Between Right Atrial Stiffness and Chamber Pressure to Quantitatively

Define Myocardial Tensile Properties.” ASME. Frontiers in Biomedical

Devices, 2017 Design of Medical Devices Conference.

doi:10.1115/DMD2017-3491

8.2 Synopsis Introduction With current leadless pacing systems, fixation is verified by applying a tensile force to the proximal end of the device. During these tests, tensile force applied to the device, causes it to displace This study aims to better define the relationships between right atrial compliance and the chamber pressures within the right atrium, so to characterize the link between tensile displacement within the right atrium, and the force exerted on an implanted device in a functional heart.

Methods Swine hearts (n=10) were obtained and reanimated on an external apparatus. A suture was attached to the right atrial appendage, passed through the heart, and

152

fixed to a force transducer. A constant velocity test was performed, inverting the right atrial appendage at a quasistatic 120mm/min, while measuring tensile load as a function of displacement. The right atrial pressure was controlled across a physiological range 0-30mmHg. Subsequently, the relative relationships between mean arterial pressure, compliance of the right atrial appendage, inversion displacement, and inversion force were derived.

Results Tensile force on the right atrium increased linearly as a function of tensile displacement, within the tested range. Right atrial stiffness and inversion force was found to linearly increase with increases in right atrial pressure:

Stiffness (N/mm) = 0.002*RAP (mmHg) + 0.02, R2 = 0.53

Inversion Force (N) = 0.08*RAP (mmHg) + 0.64, R2 = 0.52

Conclusions

This dataset quantitatively defines the relationship between tensile force and displacement within the right atrial appendage of an isolated, physiologically- representative preparation. These data may be used in the design and implementation of right atrial fixation mechanisms.

153

8.3 Background

Developing a successful cardiac device requires detailed knowledge of cardiac mechanical properties. For example, tissue failure characteristics and compliance feed into design criteria for many pacemaker leads 94.

Figure 8.1 A typical active fixation lead within the right atrial appendage. Atlas of Human Cardiac Anatomy

In the right atrium, tensile forces are exerted on the right atrial appendage in multiple clinical procedures. In a traditional lead implant, mechanical manipulations with a stylet aid a clinician in assessing lead fixation, with a seldom used “tug” test providing additional input. Atrial lead dislodgement remains one of the top complications for bradycardia pacing leads, in part because there is no standard mechanical assessment at implant to verify fixation.

Thus, a deeper understanding of forces exerted on the atrium during implant, is fundamental to understanding the problem. Further characterization of the

154

biomechanics relevant to atrial device implants will provide valuable design input for fixation tests and help drive research toward new atrial fixation mechanisms.

This study aims to better define the relationships between right atrial compliance and the chamber pressures within the right atrium, so to characterize the link between tensile displacement within the right atrium, and the force exerted on an implanted device in a functional heart. These experiments quantitatively define the fixation force of a fixed cardiac device with a given pulled displacement; i.e. displacing the device a given distance will effectively ensure the experimentally derived fixation force.

8.4 Methods

Swine hearts (n=10) were obtained and reanimated on an external apparatus, using previously described Visible Heart® methodologies 89. A suture was threaded through the epicardial surface of the right atrial appendage, as depicted in Figure 8.2A. The suture was threaded through the orifice of the tricuspid valve, ensuring minimal frictional contact with the valvular apparatus.

The suture subsequently passed through a 24 French hemostasis port placed near the right ventricular apex, across a pulley, terminating attachment to a force transducer; see Figure 8.2B.

155

A constant velocity test was performed, inverting the right atrial appendage at a quasistatic 120mm/min, while measuring tensile load as a function of displacement. Mean right atrial pressure was recorded via a pressure transducing catheter placed through the superior vena cava and into the superior-medial portion of the right atrium.

156

The mean right atrial pressure was controlled across a physiological range

0-30 mmHg by mechanical alterations in both right sided preloads and afterloads.

Subsequently, the relative relationships between mean arterial pressure, compliance of the right atrial appendage, inversion displacement, and inversion force were derived.

Figure 8.3 Example Force vs Displacement Curve

157

8.5 Results

As depicted in Figure 8.4, the relationship between mean right atrial pressure and right atrial appendage compliance was linear. The representative equation for this relationship was:

Stiffness (N/mm) = 0.002*RAP (mmHg) + 0.02

R2 = 0.53

Figure 8.4 Right Atrial Stiffness vs RA Chamber Pressure

158

Furthermore, the relationship between mean right atrial pressure and peak force at inversion of the right atrial appendage was also estimated to be linear.

Specifically, the relationship was determined to be:

Inversion Force (N) = 0.08*RAP (mmHg) + 0.64

R2 = 0.52

Both relationships show stronger intra-animal linearity than population linearity (R2 > 0.65 within individuals for compliance R2 > 0.70 within individuals for inversion force).

Figure 8.5 Right Atrial Inversion Force vs RA Chamber Pressure

Figure 8.5: Right atrial force of inversion varied with right atrial pressure. As mean RAP increases, inversion force increased approximately linearly.

159

8.6 Discussion

We have developed a novel benchtop testing setup to more critically characterize the mechanical tissue properties of the right atrium in a viable, contracting reanimated swine heart preparation. Uniquely, we were able to modulate right atrial pressure to assess acute biomechanical properties under a variety of physiologic conditions. This process is currently being extended to include human samples, thus driving translational impact.

We observed that decrease in atrial compliance corresponded with heightened right atrial pressures—i.e. as pressure increases, it becomes more difficult to pull the right atrium with a given displacement. It follows that in a patient with an increased central venous pressure, a shorter displacement pull-and-hold test should be considered to ensure adequate device fixation. This dataset quantitatively defines the relationship between tensile force and displacement within the right atrial appendage of an isolated, physiologically-representative preparation.

160

Section IV. End-of-Life Consideration for Leadless Pacemakers

One of the largest questions facing up-and-coming leadless pacing technologies is the decision of what to do once the device has reached the end of its life.

Unlike transvenous pacemakers, leadless pacemakers are implanted entirely within the cardiac chambers, resulting in an increased difficulty of explant at term.

This section of my thesis explores the extraction of leadless pacemakers, focusing on the unique challenges that have arisen in attempts to extract the

Medtronic MicraTM leadless pacemakers.

161

Chapter 9: Retrieval of a Chronically Implanted

Leadless Pacemaker Within an Isolated Heart using

Direct Visualization

Pierce J. Vatterott, MD1, Michael D. Eggen, PhD2,4, Alexander R. Mattson,

BS2,3,4, Pamela K. Omdahl, MBA4, Kathryn E. Hilpisch, BS4, Paul A. Iaizzo, PhD,

FHRS2,3

1United Heart and Vascular Clinic, St. Paul, MN USA

2Department of Surgery, University of Minnesota, Minneapolis, MN USA

3Department of Biomedical Engineering, University of Minnesota, Minneapolis,

MN USA

4Medtronic, Mounds View, MN USA

162

9.1 Preface

Perhaps the most significant question facing the leadless pacemaker are end-of-life considerations. Current leadless pacing systems have ~10 years of battery life. Once the battery expires, a clinician must make a decision on what to do with the device, often abandonment34 or retrieval.95 In the case of the

Medtronic MicraTM, retrieval attempts may utilize the MicraTM TPS delivery system in combination with a 7mm snare. In this chapter, we provide step-by- step visualization of the retrieval process in a chronically implanted animal, to highlight how some of the challenges in the procedure may be overcome.

Pierce Vatterott, Michael Eggen, and Kathryn Hilpisch contributed to study conception, design, and implementation. All listed authors contributed to manuscript preparation and editing. The following chapter has been published as an original manuscript in HeartRhythm Case Reports:

Vatterott PJ, Eggen MD, Mattson AR, Omdahl PK, Hilpisch KE, Iaizzo PA.

Retrieval of a chronically implanted leadless pacemaker within an isolated

heart using direct visualization. HeartRhythm Case Reports.

2018;4(5):167-169. doi:10.1016/j.hrcr.2017.11.014.

163

9.2 Synopsis Introduction Shown here is the retrieval of a chronically implanted leadless pacemaker in a reanimated swine heart. A superior retrieval approach was utilized and observed using direct visualization. Analysis of this multimodal imaging allows for a better understanding of the retrieval procedure and device-tissue interactions.

Case Report A Micra™ single-chamber transcatheter pacing system was implanted within the right ventricular apex of a Yorkshire-Cross swine (75.2 kg). The Micra™ device was enveloped in a resorbable meshed coating prior to implant, to promote rapid encapsulation. Eight weeks post-implant, the animal’s heart was recovered and reanimated with a clear Krebs-Henseleit buffer; internal anatomy and extraction procedures were viewed via endoscopic cameras. Employing a superior approach via the superior vena cava, a foreign body single-loop retrieval snare was used in conjunction with the MicraTM TPS delivery catheter to extract the chronicallyl implanted leadless pacemaker.

Discussion A chronically implanted MicraTM leadless pacemaker was successfully retrieved from an isolated swine heart. In this example, the pacemaker was not fully encapsulated even at 8 weeks post-implant, and thus the snare could be placed tightly around the proximal retrieval feature. Although the techniques used in this report were used to remove a clinically implanted MicraTM device at 406 days, it is unknown what the level of encapsulation was on the device, and what the overall success rate will be with this technique in fully encapsulated devices. 164

Conclusions Although retrieval of a leadless pacemaker using direct visualization is not representative of the visualization techniques available in a clinical setting, the images presented here have notable educational value for both clinicians and design engineers.

165

9.3 Introduction Recent outcomes associated with the use of leadless pacemakers have shown reduced or minimal chronic complications compared with traditional leaded systems, e.g., lead infections, pocket infections, and/or lead fractures.

Yet, even with fewer system complications, there have been some described circumstances in which leadless pacemaker retrievals have been performed, including elevated pacing thresholds, need for alternate therapy, pacemaker syndrome, or prosthetic valve endocarditis.1 Shown here is the retrieval of a chronically implanted leadless pacemaker in a reanimated swine heart. A superior retrieval approach was utilized and observed using direct visualization.

Analysis of this multimodal imaging allows for a better understanding of the retrieval procedure and device-tissue interactions.

9.4 Case Report

This research followed the guidelines established in the Guide for the

Care and Use of Laboratory Animals. A Micra™ single-chamber transcatheter pacing system (TPS; Medtronic, Minneapolis, MN, USA) was implanted within the right ventricular apex of a Yorkshire-Cross swine (75.2 kg). The Micra™ device was enveloped in a resorbable meshed coating (Vicryl, Ethicon,

Somerville, NJ, USA) prior to implant, to promote rapid encapsulation. Eight weeks post-implant, the animal’s heart was recovered and reanimated with a

166

clear Krebs-Henseleit buffer, using previously described Visible Heart® methodologies.2 Endoscopic cameras (IplexFX, Olympus Corporation, Tokyo,

Japan) were placed within the right atrium and right ventricle, while continuous fluoroscopy was recorded (Ziehm Vision R, Ziehm Imaging, Nuremberg,

Germany). A A

Employing a superior approach via the superior vena cava, a foreign body single-loop retrieval snare (7mm Amplatz GooseNeck Snare,

Covidien Medtronic, Plymouth, MN, USA) was B advanced through the tether lumen of the Micra™

TPS delivery system. The delivery system was then placed through the right atrium, across the tricuspid valve, and into the right ventricular apex of the heart. The snare was advanced to the C proximal retrieval feature of the Micra™. After securing the snare to the proximal retrieval feature, the recapture cone of the delivery system was advanced, and the snare tensioned to ensure Figure 9.1 : (A) Encapsulation is proper alignment with the retrieval tool. The visualized at the distal end of the MicraTM, (B) The MicraTM is Micra™ was subsequently pulled within the retrieved, (C) Post-extraction, the fibrous capsule remains.

167

delivery system using the device cup as counter-traction. The captured device and delivery system were then removed from the heart. See Figure 9.1 and

Video Clip in Supplementary Material (Supplemental Video 9—Retrieval of a

Chronically Implanted Leadless Pacemaker).

9.5 Discussion and Conclusions

Using the delivery catheter and a snare, a chronically implanted MicraTM leadless pacemaker was successfully retrieved from an isolated swine heart using a superior approach, which offered ease in snaring the device and, once snared, allowed the surgeon to directly and effectively apply traction and counter- traction. In this example, the pacemaker was not fully encapsulated even at 8 weeks post-implant, and thus the snare could be placed tightly around the proximal retrieval feature. During retrieval, the distal encapsulated tissue separated from the device body as it was pulled into the device cup, and this tissue remained attached to the endocardium forming a small “sock” (Figure

9.1C). It is important to note that it is unknown whether these methodologies and tools are applicable for extraction of a fully encapsulated device that could potentially be composed of calcified tissue.

As such, any required retrieval of MicraTM devices will not be limited to the techniques presented in this report. Indeed, snaring the proximal retrieval feature

168

of the MicraTM device with the compatible 7mm snare may not be practical in all clinical scenarios (for example, where tissue covering the retrieval feature exceeds 7mm in diameter). In these cases, larger snares may be used in conjunction with a steerable sheath. To date, there is limited data reported for

MicraTM retrievals. There have been five percutaneous retrieval attempts reported

(implant time range 5-406 days), with successful removal of 3 out of 5 devices.

For the 3 successful removals, 2/3 used a deflectable sheath and a larger loop snare without counter-traction (5 and 16 days), and one removal (406 days) used the techniques demonstrated in this study.3 Although the techniques reported here were used to remove a MicraTM device at 406 days, it is unknown what the level of encapsulation was on the device, and what the overall success rate will be with this technique in fully encapsulated devices. As such the technique shown in this study is primarily intended for early retrieval cases where the proximal retrieval feature is still accessible. The flexibility of the nitinol tines has facilitated the ability to retrieve the device without counter-traction.5

Of note, there are relatively robust extraction data on the Nanostim leadless pacemaker. Nanostim retrieval success was reported as 90.4% (66 of

73 attempts; implant duration range, 1 day to 4.0 years). Also, the reported rates of retrieval success were 86%, 93%, and 90% at <1 year, 1-2 years, and >2 years from implant, respectively. As such, the Nanostim data cannot be

169

extrapolated to the MicraTM leadless pacemaker as the fixation mechanisms and device length are not similar.

In our case study, direct visualization aided in the snaring and retrieval of the device in a healthy swine heart, where there was no interaction with trabeculation or any valve structures. Snaring and retrieval may be more difficult when only utilizing fluoroscopy as in a clinical setting, and also when considering the high levels of trabeculation occurring within the human heart. Although retrieval of a leadless pacemaker using direct visualization is not representative of the visualization techniques available in a clinical setting, the images presented here have notable educational value for both clinicians and design engineers.

170

Chapter 10: Atraumatic Extrication of Leadless

Pacemaker Tines from the Tricuspid Valve Apparatus

Alexander R. Mattson, BS1,2, Jorge Zhingre-Sanchez, BS1, Paul A. Iaizzo, PhD1

1Departments of Biomedical Engineering and Surgery, University of Minnesota,

Minneapolis, MN USA

2Medtronic, Mounds View, MN USA

171

10.1 Preface As described in the previous chapter, retrievals of the Medtronic MicraTM have been performed using the standard MicraTM delivery system paired with a

7mm snare. However, in practice, many retrievals of the MicraTM pacemaker have utilized larger diameter-off the shelf snares, which are not compatible with the MicraTM delivery system. In these extractions, the MicraTM device is removed from tissue, and pulled across the tricuspid valve as it exits the heart. As the steerable sheaths used in these procedures do not have a lumen size large enough to accommodate the body of the MicraTM, the fixation tines are exposed during the retrieval. This study tests whether the fixation tines run the risk of damaging the tricuspid valve apparatus during these extraction procedures.

Jorge Zhingre Sanchez contributed to data collection, experimental design, manuscript preparation and editing. Paul Iaizzo contributed to manuscript preparation and editing.

The analysis provided in this chapter was originally published as an abstract at the HeartRhythm Society 2018 Annual meeting. The full chapter contents have been submitted as a manuscript to Pacing and Clinical

Electrophysiology.

10.2 Synopsis Introduction

As there is today, no manufacturer-supplied retrieval tool for the MicraTM pacemaker (Medtronic Inc, Minneapolis, MN), off-the-shelf catheters have been employed for retrievals. In such, the proximal retrieval feature of the MicraTM is 172

snared, and the device is retracted from the myocardium, pulling the device through the tricuspid valve.

Objective

This study characterizes the potential risks of MicraTM nitinol tine engagement with the tricuspid sub-valvular apparatus.

Methods

Fresh human hearts (N=6, non-viable for transplant), were obtained from our local organ procurement agency (LifeSource, Minneapolis, MN). MicraTM fixation tines were affixed to a linear force transducer. Tines were engaged in tricuspid chordae tendineae and underwent a constant velocity tensile test. Each test was run until tines disengaged from the chordae tendineae, or ruptured tissue of the valve apparatus. Subsequently, biomechanical failure properties of the valve apparatus and isolated chordae tendineae were determined using a series of uniaxial tensile tests.

Results

There were no chordal ruptures observed during the MicraTM tine extraction testing (0/100 tests). Chordal failure required 15.0x the force of extracting a single engaged tine, and 9.0x the force of extracting two engaged tines. The uniaxial stress required for isolated chordal failure averaged 17.1N/mm2; failure strains exceed 150% resting chordal length.

173

Conclusions

The forces required to rupture tricuspid chordae tendineae significantly exceed the forces potentially imposed on the chordae during MicraTM device extractions.

We conclude that the fixation tines of the MicraTM device are unlikely to damage the tricuspid apparatus during either implant or extraction.

174

10.3 Introduction

Leadless pacing systems have shown reduced chronic complication rates compared with traditional leaded systems, eliminating the risk of lead infections, pocket infections, and/or lead fractures. Yet, even with fewer system complications, there have been some described circumstances in which leadless pacemaker retrievals have been performed, including elevated pacing thresholds, need for alternate therapy, pacemaker syndrome, or prosthetic valve endocarditis.

There are two leadless pacing systems available for clinical practice: the

Abbott NanostimTM and the Medtronic MicraTM. The NanostimTM device has a relatively robust history of short-term chronic retrieval, after battery issues necessitated the removal of many implanted devices.37,38 Despite the comparatively modest extraction history, there are reports of clinical retrieval of the Medtronic MicraTM .41 Although current retrieval experiences of the MicraTM leadless pacemaker are limited, the flexibility of the nitinol tines has facilitated the ability to retrieve the device without counter-traction.35

There is no manufacturer-supplied chronic retrieval tool for the MicraTM pacemaker. As such, devices may be retrieved using the standard MicraTM delivery catheter, paired with a foreign body, single-loop retrieval snare (7mm

Amplatz GooseNeck Snare, Covidien Medtronic, Plymouth, MN, USA). In other 175

retrieval cases, off-the-shelf catheters have been used for retrieval.96 In these cases, the proximal retrieval feature of the MicraTM device is snared, and the device is removed from the myocardium without the use of counter traction.

When using an off-the shelf catheter for extraction, the proximal retrieval feature of the MicraTM device is grasped, tension is applied, and the device fixation tines are extricated from myocardial tissue. Subsequently, the fixation tines are pulled through the tricuspid valve annulus, into the introducer, and removed from the body (Figure 10.1). As the fixation tines are pulled across the tricuspid annulus, there is a potential of tine interaction and entanglement with the tricuspid sub- valvular apparatus. It is unknown whether leadless pacemaker fixation tines might interfere with or damage the tricuspid valve apparatus, specifically the chordae tendineae, during extraction.

176

10.4 Objective

This study defines the potential risks of nitinol tine engagement with the tricuspid sub-valvular apparatus. We provide biomechanical evidence to show that engaged tines have a large safety factor to prevent damage of the valve tissues.

10.5 Methods

Human (n=10) hearts, deemed non-viable for transplant, were obtained from our local procurement agency (LifeSource, Minneapolis, MN): donor patient histories are detailed in Table 10.1. The hearts were sectioned along the right ventricular outflow tract to expose an in-tact tricuspid valve. Leadless pacemaker fixation tines (MicraTM, Medtronic, plc) were mounted to a rigid rod and affixed to a linear force transducer (Chatillon, Ametek Inc). Tines were engaged with individual tricuspid valve chordae tendineae (Figure 10.2A). As the four fixation tines are distributed evenly around the circumference of the tip of the MicraTM device, it was assumed that up to two tines could be engaged with an individual chordae tendineae at a single time. As such, extrication testing was performed in two configurations: 1) with a single fixation tine engaged, and 2) with two fixation tines engaged on the same chordus.

Chordae, with engaged tines, underwent a 120 mm/min constant velocity tensile test, measuring tensile force as a function of displacement. Each

177

extrication test was run until the fixation tines disengaged or tissue failure occurred.

A B

Two additional testing paradigms were undertaken to characterize true failure forces of the chordal complexes. First, to measure failure forces of the entire chordal complex (i.e. including the tricuspid leaflet, papillary muscle attachments, and surrounding myocardium), in an in-vivo loading analog, a fine wire three-point bending test was performed. A rigid rod was threaded beneath a single chordus and vertically displaced at a constant 120 mm/min until failure of either the papillary muscle, valve leaflet, or chordae tendineae occurred. The location of failure was noted (e.g. central chordae tendineae tear, leaflet tear, papillary muscle tear). 178

Table 10.2: Medical Histories for Donated Specimens

Heart Number Age Gender Medical History

436 54 F COPD

438 51 F Hypertension, Hyperthyroidism

450 59 F Hemorrhagic stroke, Chronic lung disease

453 48 F Hemorrhagic stroke

457 78 F Hemorrhagic stroke

458 62 M Hemorrhagic stroke

472 52 M Ischemic stroke, Hypertension

480 56 M Hemorrhagic stroke

483 50 F Hemorrhagic stroke, Hypertension

493 56 M Hemorrhagic stroke

Subsequently, to isolate biomechanical properties of the chordae tendineae themselves, a uniaxial tension testing test was performed on human

(n=4) hearts. Chordae tendineae from the tricuspid valve were dissected, preserving leaflet and papillary tissues at the chordal junction points. Custom grips fixed the junction points at either end of the chordus. Each chordus was

(Figure 10.2B) its leaflet attachment type (primary vs secondary) and papillary complex attachment (anterior, posterior, and/or septal). Uniaxial tension was

179

applied to the isolated chordae in order to determine the chordal failure stress and the location within the chordal body most susceptible to rupture. For each uniaxial pull, force was recorded as a function of tensile displacement, and rupture forces were normalized to cross-sectional areas of each sample. The location of failure or rupture (near leaflet, near papillary, or in the chordal body) was documented.

10.6 Results

The forces required to rupture the chordae tendineae, or associated leaflet/papillary complex junction points greatly exceeded the forces required to extract device fixation tines. Notably, during 100 runs of extrication testing, no ruptures of the chordal complex were observed. In testing experiments with both one and two fixation tines engaged with a single chordus, extrication was performed without tissue damage. Further, as shown in Figure 10.3, the forces required to rupture chordae tendineae or the associated valve leaflet/papillary complex greatly exceed the forces required to atraumatically extricate the fixation tines, when either one (Failure force exceeds by 15.0x) or two (Failure force exceeds by 9.0x) tines were engaged with the individual chordae.

180

Forces of chordal failure varied depending on the anatomic attachment of the chordae tendineae being tested. As shown in Figure 10.5, chordae tendineae attaching to either the posterior or septal complexes of the tricuspid valve required significantly less force to rupture than those attaching to the anterior papillary complex.

181

Chordae tendineae ruptures were binned into three categories, dependent on where tissue failure occurred: leaflet attachment/leaflet body, papillary attachment/papillary body, or within the body of the chordae tendineae. Chordae most commonly failed at the junction of the chordus with the tricuspid valve leaflet. Very rarely was failure seen within the body of the chordae tendineae

(Figure 10.5).

182

Uniaxial tensile testing of isolated chordae tendineae produced failure stresses and strains of the chordal body itself. In this testing, tricuspid chordae failed at an average stress of 17.1 N/mm2. Septal complex chordae exhibited significantly higher rupture stresses (24.9 N/mm2) than anterior (13.7 N/mm2) and posterior (14.1 N/mm2) complex chordae (Figure 10.6). In parallel, septal complex chordae showed significantly larger failure strains (288%) compared to the anterior (164%) and posterior (185%) chordae. For all the fresh tricuspid

183

chordae tested, approximately 53% ruptured near the leaflet compared to 24% near the papillary end, and 24% in the middle of the chordal body.

10.7 Discussion

Currently, there is no clinically-available extraction tool suited for the

MicraTM transcatheter pacing system. As such, some retrieval attempts involve extricating the device using an off-the-shelf retrieval catheter and foreign body retrieval snare. In such cases, the MicraTM device is removed from the myocardium, and pulled through the tricuspid valve annulus unsheathed. While crossing the tricuspid annulus, there is the potential for tine engagement in the tricuspid sub-valvular apparatus, particularly in the chordae tendineae.

Despite their fragile appearance, we have shown that the tricuspid chordae tendineae are robust structures, capable of withstanding a significant

184

amount of tensile force. Here, we found the average uniaxial failure stress of the chordae tendineae to be 17.1 N/mm2.

Notably, the forces applied by an engaged MicraTM fixation tine are significantly smaller than the failure forces noted above. Indeed, atraumatic extrication was achieved in all 100 tests where the device fixation tines were engaged with tricuspid chordae tendineae (50 tests with a single tine engaged,

50 tests with two tines engaged). We find that the forces required to rupture the chordal complex exceed the forces of tine extrication by a factor of 15.0x and

9.0x for single and two tine engaged scenarios, respectively. Uniaxial pulls on isolated chordae samples were performed to further validate the robustness of the chordae tendineae, as only a minor percentage (1.7%) of three-point tensile test failures occurred in the chordal body. Here, we find that significant stress and strain are required to rupture an isolated chordae tendineae. As such, damage to the tricuspid apparatus is unlikely to occur due to tine entanglement during a MicraTM extraction.

Further validating the reported biomechanical evidence, we have provided supplemental footage (Supplemental Video 10—Atraumatic Tine Extrication from the Tricuspid Valve) of Micra tines atraumatically extracting from the tricuspid apparatus in an isolated, human heart model, reanimated and imaged using

Visible Heart® methodologies.89 Post-tine removal, an undamaged tricuspid valve apparatus may be seen.

185

It is important to note that this dataset relates only to the interaction between the MicraTM fixation tines and the tricuspid sub valvular apparatus.

During a chronic extraction of a leadless pacemaker, damage to the tricuspid valve apparatus may be incurred through other mechanisms. Devices implanted along the high septum may have chronic encapsulation profiles that include portions of the tricuspid valve apparatus. Additionally, the use of retrieval catheters and snares could negatively interact with the tricuspid valve.

10.8 Conclusion

Here, we have provided substantial biomechanical evidence that there is minimal risk of leadless pacemaker tine entanglement damaging the tricuspid valve apparatus during a MicraTM leadless device extraction. Failure forces of the tricuspid valve apparatus exceed the forces of atraumatic tine extrication by a factor of 15.0x.

186

Thesis Summary This thesis provides several key building blocks for guiding design and implementation of next-generation pacing systems. We have quantified the anatomies which offer potentially advantageous pacing substrate for next- generation leadless pacing systems and epicardial pacemakers to target.

Additionally, we have described potentially optimized locations for physiologic pacing through the bundle of His.

The biomechanics work described in this thesis builds the foundation for the design and implementation of next-generation leadless pacing fixation in the right atrium. Notably, we have defined key properties to aid designers in avoiding perforation of the thin-walled right atrium.

Finally, this thesis examined some of the key challenges in deciding the end-of-life scenario for leadless pacing systems. Here, we describe and visualize an extraction methodology using the MicraTM TPS delivery system. We also show that extractions utilizing off-the-shelf catheters have a low risk of damaging the tricuspid apparatus, a key consideration in retrievals of the MicraTM leadless pacemaker.

The research in this thesis will help expand pacemaker technology to better serve the more than 600,000 patients implanted with permanent pacing systems each year.

187

Section V. Bibliography

1. Mond HG, Proclemer A. The 11th world survey of cardiac pacing and implantable cardioverter-defibrillators: calendar year 2009--a World Society of Arrhythmia’s project. Pacing Clin Electrophysiol. 2011;34(8):1013-1027. doi:10.1111/j.1540-8159.2011.03150.x. 2. Lidwill MC. Cardiac Disease in Relation to Anaesthesia. In: Transactinos of the Third Session, Australasian Medical Congress (British Medical Association). ; 1929:160-163. 3. HYMAN AS. Resuscitation of the stopped heart by inracardial therapy. II. Experimental use of an artificial pacemaker. Arch Intern Med. 1932;50:283- 305. 4. Zoll PM. Resuscitation of the Heart in Ventricular Standstill by External Electric Stimulation. N Engl J Med. 1952;247(20):768-771. doi:10.1056/NEJM195211132472005. 5. Iaizzo PA, Mahre MA. Historical Perspective of Cardiovascular Devices and Techniques Associated with the University of Minnesota. In: Handbook of Cardiac Anatomy, Physiology, and Devices. Cham: Springer International Publishing; 2015:439-456. doi:10.1007/978-3-319-19464- 6_25. 6. Bakken EE. One Man’s Full Life. Medtronic; 1999. 7. Lillehei CW, Gott VL, Hodges PC, Long DM, Bakken EE. TRANSISTOR PACEMAKER FOR TREATMENT OF COMPLETE ATRIOVENTRICULAR DISSOCIATION. J Am Med Assoc. 1960;172(18):2006. doi:10.1001/jama.1960.03020180016003. 8. Altman LK. Arne H. W. Larsson, 86; Had First Internal Pacemaker. The New York Times. January 18, 2002. 9. CHARDACK WM, GAGE AA, GREATBATCH W. A transistorized, self- contained, implantable pacemaker for the long-term correction of complete heart block. Surgery. 1960;48:643-654. http://www.ncbi.nlm.nih.gov/pubmed/13692461. Accessed August 29, 2017. 10. HUNTER SW, ROTH NA, BERNARDEZ D, NOBLE JL. A bipolar myocardial electrode for complete heart block. J Lancet. 1959;79:506-508. http://www.ncbi.nlm.nih.gov/pubmed/14405471. Accessed August 30, 2017. 11. Furman S, Escher DJ, Schwedel JB, Solomon N, William-Olsson F, Burford TH. Transvenous pacing. A seven-year review. Am Heart J. 1966;71(3):408-416. doi:10.1016/0002-8703(66)90483-2. 12. Pande GS. Thermoplastic polyurethanes as insulating materials for long- 188

life cardiac pacing leads. Pacing Clin Electrophysiol. 1983;6(5 Pt 1):858- 867. http://www.ncbi.nlm.nih.gov/pubmed/6195604. Accessed September 27, 2017. 13. Mond H, Sloman G. The small-tined pacemaker lead--absence of dislodgement. Pacing Clin Electrophysiol. 1980;3(2):171-177. http://www.ncbi.nlm.nih.gov/pubmed/6160506. Accessed September 27, 2017. 14. D’Allonnes GR, Pavin D, Leclercq C, et al. Long-term effects of biatrial synchronous pacing to prevent drug-refractory atrial tachyarrhythmia: a nine-year experience. J Cardiovasc Electrophysiol. 2000;11(10):1081- 1091. http://www.ncbi.nlm.nih.gov/pubmed/11059970. Accessed September 27, 2017. 15. Cazeau S, Ritter P, Bakdach S, et al. Four chamber pacing in dilated cardiomyopathy. Pacing Clin Electrophysiol. 1994;17(11 Pt 2):1974-1979. http://www.ncbi.nlm.nih.gov/pubmed/7845801. Accessed September 27, 2017. 16. Auricchio A, Stellbrink C, Block M, et al. Effect of pacing chamber and atrioventricular delay on acute systolic function of paced patients with congestive heart failure. The Pacing Therapies for Congestive Heart Failure Study Group. The Guidant Congestive Heart Failure Research Group. Circulation. 1999;99(23):2993-3001. http://www.ncbi.nlm.nih.gov/pubmed/10368116. Accessed September 27, 2017. 17. Kass DA, Chen CH, Curry C, et al. Improved left ventricular mechanics from acute VDD pacing in patients with dilated cardiomyopathy and ventricular conduction delay. Circulation. 1999;99(12):1567-1573. http://www.ncbi.nlm.nih.gov/pubmed/10096932. Accessed September 27, 2017. 18. St John Sutton M, Cerkvenik J, Borlaug BA, et al. Effects of Cardiac Resynchronization Therapy on Cardiac Remodeling and Contractile Function: Results From Resynchronization Reverses Remodeling in Systolic Left Ventricular Dysfunction (REVERSE). J Am Heart Assoc. 2015;4(9):e002054. doi:10.1161/JAHA.115.002054. 19. Cleland JGF, Daubert J-C, Erdmann E, et al. The Effect of Cardiac Resynchronization on Morbidity and Mortality in Heart Failure. N Engl J Med. 2005;352(15):1539-1549. doi:10.1056/NEJMoa050496. 20. Abraham WT, Fisher WG, Smith AL, et al. Cardiac Resynchronization in Chronic Heart Failure. N Engl J Med. 2002;346(24):1845-1853. doi:10.1056/NEJMoa013168. 21. Deshmukh P, Casavant DA, Romanyshyn M, Anderson K. Permanent, direct His-bundle pacing: a novel approach to cardiac pacing in patients with normal His-Purkinje activation. Circulation. 2000;101(8):869-877. http://www.ncbi.nlm.nih.gov/pubmed/10694526. Accessed September 28,

189

2017. 22. Occhetta E, Bortnik M, Magnani A, et al. Prevention of ventricular desynchronization by permanent para-Hisian pacing after atrioventricular node ablation in chronic atrial fibrillation: a crossover, blinded, randomized study versus apical right ventricular pacing. J Am Coll Cardiol. 2006;47(10):1938-1945. doi:10.1016/j.jacc.2006.01.056. 23. Barba-Pichardo R, Morina-Vazquez P, Fernandez-Gomez JM, Venegas- Gamero J, Herrera-Carranza M. Permanent His-bundle pacing: seeking physiological ventricular pacing. Europace. 2010;12(4):527-533. doi:10.1093/europace/euq038. 24. ZANON F, SVETLICH C, OCCHETTA E, et al. Safety and Performance of a System Specifically Designed for Selective Site Pacing. Pacing Clin Electrophysiol. 2011;34(3):339-347. doi:10.1111/j.1540- 8159.2010.02951.x. 25. Kronborg MB, Mortensen PT, Poulsen SH, Gerdes JC, Jensen HK, Nielsen JC. His or para-His pacing preserves left ventricular function in atrioventricular block: a double-blind, randomized, crossover study. Europace. 2014;16(8):1189-1196. doi:10.1093/europace/euu011. 26. Sharma PS, Dandamudi G, Naperkowski A, et al. Permanent His-bundle pacing is feasible, safe, and superior to right ventricular pacing in routine clinical practice. Hear Rhythm. 2015;12(2):305-312. doi:10.1016/j.hrthm.2014.10.021. 27. Lustgarten DL, Crespo EM, Arkhipova-Jenkins I, et al. His-bundle pacing versus biventricular pacing in cardiac resynchronization therapy patients: A crossover design comparison. Hear Rhythm. 2015;12(7):1548-1557. doi:10.1016/j.hrthm.2015.03.048. 28. Barba-Pichardo R, Manovel Sánchez A, Fernández-Gómez JM, Moriña- Vázquez P, Venegas-Gamero J, Herrera-Carranza M. Ventricular resynchronization therapy by direct His-bundle pacing using an internal cardioverter defibrillator. EP Eur. 2013;15(1):83-88. doi:10.1093/europace/eus228. 29. Madhavan M, Mulpuru SK, McLeod CJ, Cha Y-M, Friedman PA. Advances and Future Directions in Cardiac Pacemakers. J Am Coll Cardiol. 2017;69(2):211-235. doi:10.1016/j.jacc.2016.10.064. 30. Reynolds D, Duray GZ, Omar R, et al. A Leadless Intracardiac Transcatheter Pacing System. N Engl J Med. 2016;374(6):533-541. doi:10.1056/NEJMoa1511643. 31. Reddy VY, Knops RE, Sperzel J, et al. Permanent Leadless Cardiac Pacing: Results of the LEADLESS Trial. Circulation. 2014;129(14):1466- 1471. doi:10.1161/CIRCULATIONAHA.113.006987. 32. Udo EO, Zuithoff NPA, van Hemel NM, et al. Incidence and predictors of short- and long-term complications in pacemaker therapy: The FOLLOWPACE study. Hear Rhythm. 2012;9(5):728-735.

190

doi:10.1016/j.hrthm.2011.12.014. 33. Kirkfeldt RE, Johansen JB, Nohr EA, Jorgensen OD, Nielsen JC. Complications after cardiac implantable electronic device implantations: an analysis of a complete, nationwide cohort in Denmark. Eur Heart J. 2014;35(18):1186-1194. doi:10.1093/eurheartj/eht511. 34. Omdahl P, Eggen MD, Bonner MD, Iaizzo PA, Wika K. Right Ventricular Anatomy Can Accommodate Multiple Micra Transcatheter Pacemakers. PACE - Pacing Clin Electrophysiol. 2016;39(4):393-397. doi:10.1111/pace.12804. 35. Eggen MD, Grubac V, Bonner MD. Design and Evaluation of a Novel Fixation Mechanism for a Transcatheter Pacemaker. IEEE Trans Biomed Eng. 2015;62(9):2316-2323. 36. Tjong FVY, Reddy VY. Permanent Leadless Cardiac Pacemaker Therapy: A Comprehensive Review. Circulation. 2017;135(15):1458-1470. doi:10.1161/CIRCULATIONAHA.116.025037. 37. Lakkireddy D, Knops R, Atwater B, et al. A worldwide experience of the management of battery failures and chronic device retrieval of the Nanostim leadless pacemaker. Hear Rhythm. July 2017. doi:10.1016/j.hrthm.2017.07.004. 38. Reddy VY, Miller MA, Knops RE, et al. Retrieval of the Leadless Cardiac Pacemaker: A Multicenter Experience. Circ Arrhythmia Electrophysiol. 2016;9(12). doi:10.1161/CIRCEP.116.004626. 39. Karim S, Abdelmessih M, Marieb M, Reiner E, Grubman E. Extraction of a Micra Transcatheter Pacing System: First-in-human experience. Hear Case Reports. 2016;2(1):60-62. doi:10.1016/j.hrcr.2015.10.001. 40. P Vatterott, MD Eggen, AR Mattson, P Omdahl, K Hilpisch PI. Retrieval of a Chronically Implanted Leadless Pacemaker within an Isolated Heart using Direct Visualization. Hear Rhythm Case Reports. 2018. 41. Grubman E, Ritter P, Ellis CR, et al. To retrieve, or not to retrieve: System revisions with the Micra transcatheter pacemaker. Hear Rhythm. July 2017. doi:10.1016/j.hrthm.2017.07.015. 42. Reddy VY, Miller MA, Neuzil P, et al. Cardiac Resynchronization Therapy With Wireless Left Ventricular Endocardial Pacing: The SELECT-LV Study. J Am Coll Cardiol. 2017;69(17):2119-2129. doi:10.1016/j.jacc.2017.02.059. 43. Hauser RG. The subcutaneous implantable cardioverter-defibrillator: should patients want one? J Am Coll Cardiol. 2013;61(1):20-22. doi:10.1016/j.jacc.2012.07.069. 44. Tjong FVY, Brouwer TF, Smeding L, et al. Combined leadless pacemaker and subcutaneous implantable defibrillator therapy: feasibility, safety, and performance. Europace. 2016;18(11):1740-1747. doi:10.1093/europace/euv457. 45. Tjong FVY, Brouwer TF, Kooiman KM, et al. Communicating Antitachycardia Pacing-Enabled Leadless Pacemaker and Subcutaneous

191

Implantable Defibrillator. J Am Coll Cardiol. 2016;67(15):1865-1866. doi:10.1016/j.jacc.2016.02.039. 46. Garrigue S, Jaïs P, Espil G, et al. Comparison of chronic biventricular pacing between epicardial and endocardial left ventricular stimulation using Doppler tissue imaging in patients with heart failure. Am J Cardiol. 2001;88(8):858-862. http://www.ncbi.nlm.nih.gov/pubmed/11676947. Accessed September 28, 2017. 47. Sohal M, Shetty A, Niederer S, et al. Mechanistic insights into the benefits of multisite pacing in cardiac resynchronization therapy: The importance of electrical substrate and rate of left ventricular activation. Hear Rhythm. 2015;12(12):2449-2457. doi:10.1016/j.hrthm.2015.07.012. 48. Leclercq C, Gadler F, Kranig W, et al. A Randomized Comparison of Triple-Site Versus Dual-Site Ventricular Stimulation in Patients With Congestive Heart Failure. J Am Coll Cardiol. 2008;51(15):1455-1462. doi:10.1016/j.jacc.2007.11.074. 49. Behar JM, Bostock J, Ginks M, et al. Limitations of chronic delivery of multi-vein left ventricular stimulation for cardiac resynchronization therapy. J Interv Card Electrophysiol. 2015;42(2):135-142. doi:10.1007/s10840-014- 9971-2. 50. Pappone C, Ćalović Ž, Vicedomini G, et al. Multipoint left ventricular pacing improves acute hemodynamic response assessed with pressure-volume loops in cardiac resynchronization therapy patients. Hear Rhythm. 2014;11(3):394-401. doi:10.1016/j.hrthm.2013.11.023. 51. Turakhia MP, Cao M, Fischer A, et al. Reduced Mortality Associated With Quadripolar Compared to Bipolar Left Ventricular Leads in Cardiac Resynchronization Therapy. JACC Clin Electrophysiol. 2016;2(4):426-433. doi:10.1016/j.jacep.2016.02.007. 52. Zurbuchen A, Haeberlin A, Bereuter L, et al. The Swiss approach for a heartbeat-driven lead- and batteryless pacemaker. Hear Rhythm. 2017;14(2):294-299. doi:10.1016/j.hrthm.2016.10.016. 53. Lu B, Chen Y, Ou D, et al. Ultra-flexible Piezoelectric Devices Integrated with Heart to Harvest the Biomechanical Energy. Sci Rep. 2015;5(1):16065. doi:10.1038/srep16065. 54. Haeberlin A, Zurbuchen A, Walpen S, et al. The first batteryless, solar- powered cardiac pacemaker. Hear Rhythm. 2015;12(6):1317-1323. doi:10.1016/j.hrthm.2015.02.032. 55. Rosen MR. Gene Therapy and Biological Pacing. Phimister EG, ed. N Engl J Med. 2014;371(12):1158-1159. doi:10.1056/NEJMcibr1408897. 56. Protze SI, Liu J, Nussinovitch U, et al. Sinoatrial node cardiomyocytes derived from human pluripotent cells function as a biological pacemaker. Nat Biotechnol. 2016;35(1):56-68. doi:10.1038/nbt.3745. 57. Boersma L, Barr C, Knops R, et al. Implant and Midterm Outcomes of the Subcutaneous Implantable Cardioverter-Defibrillator Registry. J Am Coll

192

Cardiol. 2017;70(7):830-841. doi:10.1016/j.jacc.2017.06.040. 58. Burke MC, Marcovechio A, Sanghera R, Ebner A. Abstract 20697: An Intracostal Pacing System: A Novel Alternative to Leadless Pacing. Circulation. 2017;136(Suppl 1). http://circ.ahajournals.org/content/136/Suppl_1/A20697. Accessed June 24, 2018. 59. Sholevar DP, Tung S, Kuriachan V, et al. Feasibility of extravascular pacing with a novel substernal electrode configuration: The Substernal Pacing Acute Clinical Evaluation study. Hear Rhythm. 2018;15(4):536-542. doi:10.1016/J.HRTHM.2017.11.030. 60. D’Avila A, Houghtaling C, Gutierrez P, et al. Catheter ablation of ventricular epicardial tissue: A comparison of standard and cooled-tip radiofrequency energy. Circulation. 2004;109(19):2363-2369. 61. Shirani J, Berezowski K, Roberts WC. Quantitative measurement of normal and excessive (cor adiposum) subepicardial adipose tissue, its clinical significance, and its effect on electrocardiographic QRS voltage. Am J Cardiol. 1995;76(5):414-418. doi:10.1016/S0002-9149(99)80116-7. 62. Rabkin SW. Epicardial fat: properties, function and relationship to obesity. Obes Rev. 2007;8(3):253-261. doi:10.1111/j.1467-789X.2006.00293.x. 63. Tansey DK, Aly Z, Sheppard MN. Fat in the right ventricle of the normal heart. Histopathology. 2005;46(1):98-104. doi:10.1111/j.1365- 2559.2005.02054.x. 64. REINER L, MAZZOLENI A, RODRIGUEZ FL. Statistical analysis of the epicardial fat weight in human hearts. AMA Arch Pathol. 1955;60(4):369- 373. http://www.ncbi.nlm.nih.gov/pubmed/13258032. Accessed February 5, 2016. 65. Corradi D, Maestri R, Callegari S, et al. The ventricular epicardial fat is related to the myocardial mass in normal, ischemic and hypertrophic hearts. Cardiovasc Pathol. 2004;13(6):313-316. doi:10.1016/j.carpath.2004.08.005. 66. Picard FA, Gueret P, Laissy J-P, et al. Epicardial Adipose Tissue Thickness Correlates with the Presence and Severity of Angiographic Coronary Artery Disease in Stable Patients with Chest Pain. Devaney J, ed. PLoS One. 2014;9(10):e110005. doi:10.1371/journal.pone.0110005. 67. Eggen M, Bateman M, Iaizzo PA. Methods to Prepare Perfusion Fixed Cardiac Specimens for Multimodal Imaging: The Use of Formalin and Agar Gels. J Med Device. 2011;5(2):27539. doi:10.1115/1.3591396. 68. Sons HU, Hoffmann V. Epicardial fat cell size, fat distribution and fat infiltration of the right and left ventricle of the heart. Anat Anz. 1986;161(5):355-373. http://www.ncbi.nlm.nih.gov/pubmed/3729005. Accessed February 5, 2016. 69. Salazar J, Luzardo E, Mejías JC, et al. Epicardial Fat: Physiological, Pathological, and Therapeutic Implications. Cardiol Res Pract.

193

2016;2016:1-15. doi:10.1155/2016/1291537. 70. Becker M, Altiok E, Ocklenburg C, et al. Analysis of LV Lead Position in Cardiac Resynchronization Therapy Using Different Imaging Modalities. JACC Cardiovasc Imaging. 2010;3(5):472-481. doi:10.1016/j.jcmg.2009.11.016. 71. Doll N, Piorkowski C, Czesla M, et al. Epicardial versus Transvenous Left Ventricular Lead Placement in Patients Receiving Cardiac Resynchronization Therapy: Results from a Randomized Prospective Study. Thorac Cardiovasc Surg. 2008;56(5):256-261. doi:10.1055/s-2008- 1038633. 72. Dekker ALA., Phelps B, Dijkman B, et al. Epicardial left ventricular lead placement for cardiac resynchronization therapy: optimal pace site selection with pressure-volume loops. J Thorac Cardiovasc Surg. 2004;127(6):1641-1647. doi:10.1016/j.jtcvs.2003.10.052. 73. Boyle NG, Shivkumar K. Epicardial Interventions in Electrophysiology. Circulation. 2012;126(14). 74. Sosa E, Scanavacca M, D’Avila A, Pilleggi F. A New Technique to Perform Epicardial Mapping in the Electrophysiology Laboratory. J Cardiovasc Electrophysiol. 1996;7(6):531-536. doi:10.1111/j.1540- 8167.1996.tb00559.x. 75. Mattson A. Placeholder Reference 1. 76. Silaghi A, Piercecchi-Marti M-D, Grino M, et al. Epicardial Adipose Tissue Extent: Relationship With Age, Body Fat Distribution, and Coronaropathy. Obesity. 2008;16(11):2424-2430. doi:10.1038/oby.2008.379. 77. Okada K, Ohshima S, Isobe S, et al. Epicardial fat volume correlates with severity of coronary artery disease in nonobese patients. J Cardiovasc Med (Hagerstown). 2014;15(5):384-390. doi:10.2459/JCM.0b013e32836094da. 78. Nerlekar N, Brown AJ, Muthalaly RG, et al. Association of Epicardial Adipose Tissue and High-Risk Plaque Characteristics: A Systematic Review and Meta-Analysis. J Am Heart Assoc. 2017;6(8):e006379. doi:10.1161/JAHA.117.006379. 79. Rosito GA, Massaro JM, Hoffmann U, et al. Pericardial Fat, Visceral Abdominal Fat, Cardiovascular Disease Risk Factors, and Vascular Calcification in a Community-Based Sample: The Framingham Heart Study. Circulation. 2008;117(5):605-613. doi:10.1161/CIRCULATIONAHA.107.743062. 80. Gorter PM, van Lindert ASR, de Vos AM, et al. Quantification of epicardial and peri-coronary fat using cardiac computed tomography; reproducibility and relation with obesity and metabolic syndrome in patients suspected of coronary artery disease. Atherosclerosis. 2008;197(2):896-903. doi:10.1016/j.atherosclerosis.2007.08.016. 81. Suárez AG, Hornero F, Berjano EJ. Mathematical modeling of epicardial

194

RF ablation of atrial tissue with overlying epicardial fat. Open Biomed Eng J. 2010;4:47-55. doi:10.2174/1874120701004020047. 82. DIXIT S, NARULA N, CALLANS DJ, MARCHLINSKI FE. Electroanatomic Mapping of Human Heart: J Cardiovasc Electrophysiol. 2003;14(10):1128- 1128. doi:10.1046/j.1540-8167.2003.03138.x. 83. Desjardins B, Morady F, Bogun F. Effect of Epicardial Fat on Electroanatomical Mapping and Epicardial Catheter Ablation. J Am Coll Cardiol. 2010;56(16):1320-1327. doi:10.1016/j.jacc.2010.04.054. 84. Kawashima T, Sasaki H. A macroscopic anatomical investigation of atrioventricular bundle locational variation relative to the membranous part of the ventricular septum in elderly human hearts. Surg Radiol Anat. 2005;27(3):206-213. doi:10.1007/s00276-004-0302-7. 85. DANDAMUDI G, VIJAYARAMAN P. The Complexity of the His Bundle: Understanding Its Anatomy and Physiology through the Lens of the Past and the Present. Pacing Clin Electrophysiol. 2016;39(12):1294-1297. doi:10.1111/pace.12925. 86. Correa de Sa DD, Hardin NJ, Crespo EM, Nicholas KB, Lustgarten DL. Autopsy analysis of the implantation site of a permanent selective direct his bundle pacing lead. Circ Arrhythm Electrophysiol. 2012;5(1):244-246. doi:10.1161/CIRCEP.111.968834. 87. Vijayaraman P, Naperkowski A, Subzposh FA, et al. Permanent His Bundle Pacing: Long-Term Lead Performance and Clinical Outcomes. Hear Rhythm. December 2017. doi:10.1016/J.HRTHM.2017.12.022. 88. Mulpuru SK, Cha Y-M, Asirvatham SJ. Synchronous ventricular pacing with direct capture of the atrioventricular conduction system: Functional anatomy, terminology, and challenges. Hear Rhythm. 2016;13(11):2237- 2246. doi:10.1016/J.HRTHM.2016.08.005. 89. Chinchoy E, Soule CL, Houlton AJ, et al. Isolated Four-Chamber Working Swine Heart Model. 2000;4975(0):1607-1614. 90. Alanis J, Benitez D. TWO PREFERENTIAL CONDUCTING PATHWAYS WITHIN THE BUNDLE OF HIS OF THE DOG HEART. Jap J Physiol. 1975;25:371-385. https://www.jstage.jst.go.jp/article/jjphysiol1950/25/3/25_3_371/_pdf. Accessed January 26, 2018. 91. Kupersmith J, Krongrad E, Waldo AL. Conduction Intervals and Conduction Velocity in the Human Cardiac Conduction System Studies during Open- Heart Surgery. http://circ.ahajournals.org/content/circulationaha/47/4/776.full.pdf. Accessed January 26, 2018. 92. Hill AJ, Iaizzo PA. Comparative Cardiac Anatomy. In: Handbook of Cardiac Anatomy, Physiology, and Devices. Totowa, NJ: Humana Press; 2005:81- 91. doi:10.1007/978-1-59259-835-9_5. 93. Ghani A, Delnoy PPHM, Ramdat Misier AR, et al. Incidence of lead

195

dislodgement, malfunction and perforation during the first year following device implantation. Neth Heart J. 2014;22(6):286-291. doi:10.1007/s12471-014-0556-6. 94. Zhao X, Wenk JF, Burger M, et al. Simulation of mechanical environment in active lead fixation: effect of fixation helix size. J Biomech Eng. 2011;133(6):61006. doi:10.1115/1.4004288. 95. Grubman E, Ritter P, Ellis CR, et al. To retrieve, or not to retrieve: System revisions with the Micra transcatheter pacemaker. Hear Rhythm. 2017. doi:10.1016/j.hrthm.2017.07.015. 96. Gerdes C, Nielsen JC. Retrieval of Medtronic Micra Transcatheter Pacing System after tether removal. Europace. 2016. doi:10.1093/europace/euv444.

196