Thanks very much. It's a pleasure to be here. And I'm really looking forward to giving this talk because I've spent a lot of time writing and thinking about some of these issues. And when you think about it, most cardiac electrophysiologists spend a lot of time in planting devices. So what I'd like to do is focus on the human dynamics of cardiac pacing. And most of my talk will focus on CRT. And I want to get you to think about CRT, particularly to give you a better appreciation of CRT. Because even today in 2020, there is a substantial percentage of patients, maybe 20 to 25% of patients, who don't respond to CRT pacing. So I want you to think about it. And maybe when you come away from this lecture, you'll actually think, boy, it's really amazing that 75% of patients do respond to CRT. So I'm going to start off with some disclosures. Now, I know you've probably heard this, but it does bear repeating. And that is about 23 or 24 years ago, the NIH funded a study called the MOST study. And it was a study where patients, over 2,000 patients, underwent implantation of dual chamber pacemakers, then had their programming randomized to either DDDR or BBIR. And what you have to think about is 80% of these patients had no or minimal heart failure. They had class 1 or 2 heart failure. And then you can see the event rates based on the percent ventricular pacing over the first 30 days. And on the top two panels is patients who had DDDR pacing in the A panel. On the bottom two panels is patients who had BBIR pacing. And what you see and what we know is that basically, the lower your ventricular pacing load, that is shown in the solid line on top and the solid line on top, the less likely you were to develop congestive heart failure, the less likely you were to develop atrial fibrillation. The same thing was true for patients who were programmed BBIR. The less ventricular pacing you did, the less likely heart failure, the less likely atrial fibrillation. And of course, this observation was made again in the David trial, which is another study comparing dual chamber pacing in patients who need ICDs. This was a smaller trial because there were only 506 patients. Randomized to DDDR at 70 beats a minute or VBI at 40 beats a minute. These patients had no indication for ventricular pacing. All patients had an EF below 35 to 40%. And if you were pacing more than 40% of the time, you had a worse outcome. And this outcome here is measured as months to death or first hospitalization for heart failure. So really hard outcome. And your quality of life was better with less pacing. So now we know ventricular pacing can cause heart failure, can cause a fib, can impair quality of life. This is one of several recent studies. This is a study from the Cleveland Clinic. There's another study published from the University of Pennsylvania. Both were published about three or four years ago, and both studies looked at the incidence of pacing induced cardiomyopathy. So now we take patients who have a normal ejection fraction, undergo implantation of a dual chamber pacemaker, which is the appropriate thing to do in patients who have a normal ejection fraction. And then they look at patients who are RV pacing greater than 20% in one trial, greater than 40% in another trial. And they look at the incidence of pacing induced cardiomyopathy. Pacing induced cardiomyopathy is defined as an EF less than 40%. And what these two studies, this is only one of the two studies showed, is that the incidence of pacing induced cardiomyopathy is about 12 to 15% in patients that preserved LV function who undergo a high degree of pacing. And this incidence of about 15% develops over about four or five years. So let's talk more about pacing induced cardiomyopathy for about five or 10 minutes before we talk about the treatment of pacing induced cardiomyopathy with CRT. So one hope was that, one hope was that maybe we could pace from the right ventricle and non-apical pacing sites. Most of these studies look at RV afloat tract pacing. These are small randomized controlled clinical trials that follow up was less than 12 months. This is a study I did with Bruce Stambler now 17 years ago called the ROVA trial. And these are studies where the followup is greater than 12 months. Now, although it suggests that in fact there is a benefit from pacing at non-RV apical sites compared to RV apical sites, there's a slight benefit in terms of improvement in ejection fraction. The bottom line is that randomized controlled trials provided inconclusive or negative benefit with respect to exercise capacity, functional class, quality of life, and survival. So RV non-apical pacing sites, other than the conduction system, have pretty much been a failure in terms of important clinical endpoints. So sort of as a segue in terms of talking about other pacing sites, this is a fascinating study. This is a cross-sectional multi-center study of 178 children from 21 centers who had primarily congenital AV block. They had a structurally normal heart and they underwent permanent pacing and were studied over about 5.4 years of followup. The median age was 11.2. The pacing sites were RV afloat tract, RV lateral wall, RV apex, and RV septum. And you see from every RV pacing site, over about a five-year follow-up, there's a fall in ejection fraction. However, contrast that with LV apical pacing, LV lateral wall pacing, or LV base pacing, shown here, the base of the left ventricle, you see that the ejection fraction is stabilized. And when you look at interventricular mechanical delay, RV LV delay, you see that that either has no effect or is improved when you pace from the RV lateral wall or RV base compared to RV septal. So this actually shows you that if you pace long enough from the RV, in most patients, you're going to cause a problem over time if you look long enough, or if you wait long enough to look, or you look with sensitive methods of detecting LV dysfunction. Now this is a multiple choice question that I'm going to actually answer for you. Approximately 15 to 20 percent of patients over five years with a normal EF and a high burden of RV apical pacing are going to develop a pacing-induced cardiomyopathy. In patients who have a low EF, that number is probably twice as high. The incidence of pacing-induced cardiomyopathy is the same whether you pace from the RV apex or the outflow tract, and the misconception that it is only rarely observed in patients with congenital complete heart block is also incorrect. Now one of the interesting things is how do you define pacing-induced cardiomyopathy? And I think it's a spectrum. You have a drop in LVEF to below 50%, a drop in LVEF but baseline is still above 50%, so they may have an EF of 65%, and five years later the EF is down to 55%. So that's a 10% drop still within normal limits. Worsening of a pre-existing cardiomyopathy, heart failure symptoms, heart failure hospitalization, AFib, or a decrease in EF that's substantial, increased BNP. Those are all different parts of the spectrum, all different portions of the spectrum of pacing-induced cardiomyopathy. And this review from Sunit Mittal is an excellent review of pacing-induced cardiomyopathy. Now I want to move on and talk about programming a pacemaker before conduction system pacing. About 10 or 15 years ago it became very popular, and every company has now come out with a method to minimize ventricular pacing because ventricular pacing is bad. However, all these different modes, Managed Ventricular Pacing or MVP, SAFAR, which is now Levonova, VIP, which is a St. Jude algorithm, and RHYTHMI, which is a Boston scientific algorithm. All these algorithms are effective at decreasing ventricular pacing. However, they have led to sometimes profound AV dyssynchrony. So what one's doing is by reducing ventricular pacing, they are sacrificing AV synchrony to sometimes patients having AV intervals that are widely non-physiologic, 300, 350, 400. Not only are those AV intervals non-physiologic, but they're probably, or in some cases, they're actually prorhythmic. Now you might say, what are the benefits of minimizing ventricular pacing? Well, this is a great study. It is a meta-analysis, and it's a meta-analysis of minimizing ventricular pacing. It looks at a number of studies, and then it looks at the benefit in terms of does minimizing ventricular pacing decrease the risk of persistent or permanent AFib? Answer, no. Does it decrease all-cause hospitalization? Answer, no. Does it decrease all-cause mortality? Answer, no. So minimizing ventricular pacing actually does not do any of the things we hoped it would do when we actually began to look at multiple studies. After the initial enthusiasm, other studies with longer follow-up came, and the overall take-home message is probably not a lot of benefit in any important clinical outcome. So this slide and the subsequent slide summarize some of the metabolic changes of RV apical pacing, remodeling with asymmetric hypertrophy, histological changes in ventricular myocardium, hemodynamic changes, and mechanical function. And all these things lead to altered hemodynamics, and in fact, we think pacing-induced cardiomyopathy has many similar features to left bundle branch block. They're both cardiomyopathies caused by dyssynchrony. In fact, our group has done a lot of work with PBC-induced cardiomyopathies. That's another cardiomyopathy that's caused by dyssynchrony. So these are all dyssynchrony-induced cardiomyopathy. We'll talk a lot about left bundle branch block. We'll talk a little bit about treating pacing-induced cardiomyopathy. And for left bundle branch block cardiomyopathy, we can do CRT, or we can pace the conduction system. For pacing-induced cardiomyopathy, we can use algorithms to minimize RV pacing. But again, those don't work well. We can do conduction system pacing, or we can do CRT, and all those are important. I segue into talking about CRT by reminding you of block heart failure. It's the first and only randomized study where patients were randomized to either CRT pacemakers or CRT defibrillators versus dual chamber pacemakers and dual chamber defibrillators. It was a composite primary endpoint, and biventricular pacing did better in these patients, but it was primarily driven by changes in LV encystolic volume. And here you see that, again, this is primarily driven by urgent care visits for heart failure, but there are some criticisms of this study. In particular, there were many patients who had first gravy block and had forced ventricular pacing as opposed to minimized ventricular pacing. But the bottom line is this study remains alone in showing a benefit of IV pacing, and people have EFs of 35 to 50 percent driven by LV encystolic volume. To talk about CRT-P, I recommend you go back and read this paper. It was published 50 years ago. It is one of the most amazing papers published in the cardiovascular literature. This is a study of seven hearts that were explanted from individuals who died of various strokes and cerebral bleeds and had no prior heart disease. Their hearts were removed within 30 minutes of death, and then the electrical activity of the heart was recorded from epicardial and intramural plunge electrodes. It's quite amazing. Their hearts were actually fixed in diastolic shape by using beeswax, and then they added potassium chloride to a perfusion fluid. They looked at unipolar and bipolar electrograms. These were annotated and measured by hand, hundreds of electrograms, and then these maps were created by hand. What's even crazier is to confirm these findings in the human heart, they then went and did a couple of dog experiments to confirm this was, in fact, the appropriate electrical activation of a heart. Basically, what one sees, if one looks at the right ventricle here, the left ventricle, which is cut open, is that one sees the electrical activation of the heart begins high, up here on the paraceptal wall, about a third of the way down from the apex to the basis, where it begins. You can see the isochrones, which refer to parts of the heart that are electrically activated at the same time. If you see this color red or the darker color red, it means all those parts of the heart are being activated in the first five milliseconds, high up here in the paraceptal wall below the mitral valve attachment. Then it continues to activate through the interventricular septum, and then slowly towards the right ventricle through the moderator band here in the specialized conduction system of the heart. Later here on the lateral wall of the right ventricle, and of course, through the septum to the free wall of the left ventricle. The latest area of ventricular activation, again, is going to be the lateral wall of the left ventricle and endocardially near the anterior papillary muscle insertion of the left ventricle. Now, these slides are from the guidelines. This is a frequent area of questions for the American Board of Internal Medicine. I just show these slides. I don't want to spend a lot of time going over the indications for CRTP, but I encourage you to spend some time, read the guidelines, become familiar with the guidelines and the updated guidelines of what patients are appropriate for CRT. These slides are quite useful to know not only what patients are appropriate for CRT, but for example, a patient has class one heart failure, has a QRS duration of 140 milliseconds, and an IVCD is a class three indication, so it's absolutely contraindicated, and you should know that as well. This shows you the hemodynamics of RV apical pacing. I think it's important to know the hemodynamics. Let me make sure I didn't skip. Okay. Hemodynamics, and this looks at DPDT and negative DPDT. The take-home message is the negative effects of RV, this is RV pacing, are amplified in patients who have impaired LV function. These negative hemodynamic effects are worse in patients who have impaired LV function, and in fact, effects on relaxation and diastole, you don't even see those with RV apical pacing in normal hearts, but again, you see those in patients who have impaired LV function. This is a circular map that shows you local external work in a dog heart with RA pacing, LV free wall pacing, and RV apical pacing, and looks at fiber stress and fiber length. This curve, the area inside this curve, represents work, and you can see normal and LV free wall pacing. You have fiber length, fiber stress, you have a rectangular, a rectangular curve with a lot of work being done with both, this is conduction AAI pacing through the normal conduction system, this is through LV free wall, but with RV apical pacing, you almost have a figure of eight with relatively little work being done with RV apical pacing, and we'll get back to that in a second. This looks at the time course of myofiber strain, and this is a fascinating slide. I want you to focus your attention to the top two portions. We'll talk a little bit about hemodynamics, we'll talk a little bit about the hemodynamics, or the time course of myofiber strain, shown in the upper left here, and a stress strain relationship shown in the upper right. Now, the septum is shown in the red, and the LV free wall is shown in the blue. You see RV and LV pressure down here, and you see this pressure volume loop. In the red, okay, in the septum, we see what's called a septal flash, and that is the early systolic leftward motion of the septum, and then we see later systolic rightward motion, which is what you call on the echocardiogram paradoxical septal motion. These early contracting fibers in the septum can shorten rapidly during isovolumic contraction because the lateral wall of the left ventricle is in a relaxed state. The lateral wall of the left ventricle is in a relaxed state. This rapid early shortening is followed by systolic stretch. It's stretched due to delayed mechanical contraction of the lateral wall and premature relaxation. The septum, the septal region, show the least systolic strain within the LV, and the lateral wall shows the highest strain, and because the septum is stretching when nothing's happening in the lateral wall, you get a figure of eight shape, again, which shows a low net area because simply the septum doesn't do a lot of external work. In regions remote from the pacing site, for example, in the lateral wall of patient's left bundle branch block, loops are wide, as you can see here, and external work can be twice that what we see during synchronous ventricular activation. The total myocardial work will be the sum of the external work and the potential energy. As you can see, it's reduced substantially in the septum and left bundle branch block. In terms of further understanding of hemodynamics, this is a computer model developed by Fritz Princeton in the Netherlands, and what he did is he looks at the onset of LV free wall activation, and he tries to create a model of left bundle branch block. He does that by delaying left ventricular free wall 20 milliseconds, 40 milliseconds, later 60, 80, 100 milliseconds. Then he measures the strain pattern, the strain patterns measured in a patient before CRT and during CRT. The first thing that's quite, the red lines indicate systolic shortening, the green, the systolic pre-stretch we see in the septum, and the blue, the septal rebound stretch. What you see is with CRT, this blue rebound stretch in the septum, which is these blue thick lines, is almost completely gone when we go from no CRT to CRT. That's very important, and you see the red, the amount of work done by systolic shortening is increased dramatically. Let me summarize the hemodynamics of LV pacing. You get improved cardiac pump function expressed as a higher LVDPDT, pulse pressure, cardiac output, and injection fraction. This benefit in hemodynamics is achieved at an unchanged or even lower filling pressure. It relieves coronary artery under perfusion, that should be coronary artery under perfusion. You get a better coordination of contraction, which improves pump function with a slight decrease actually in energy consumption. CRT benefits increase with a longer application of this therapy. That is, we see benefit progressively increasing over time, including reductions in BMP, and of course, improved heart rate variability. I wanted to show this slide just to point out, this slide summarizes a new electrocardiographic criteria for left bundle branch block, both by ECHO and EKG. Here's the EKG, a terminal S and B1 notching. You want to see notching in at least one lead and slurring of the downstroke, as you can downstroke, as you can see here. QRS criteria are QRS greater than 140 milliseconds for men and less than 130 milliseconds for women. This is basically the nadir of the septal contraction occurs within 70% of the ejection phase. That's one of the ECHO criterias. The nadir of the contraction of the posterolateral wall occurs after, number three, the nadir of the posterolateral wall occurs after AV valve is closed, and that defines the end of the LV ejection phrase. Those are the ECHO definitions of left bundle branch block. Here are the EKG definitions of left bundle branch block. That's important to understand because one of the major reasons that patients do not respond to CRT therapy is the patients don't have the right type of electrical disease. What do I mean by right type of electrical disease? Well, I think this is pretty important stuff. This data comes from a variety of sources, but it comes from endocardial mapping of patients who have conduction system disease with electroanatomic mapping systems. More recently, data comes from body surface mapping. These are endocardial maps of both RV and LV activation in three patients who have left bundle branch block. What you see is their QRS. What you see is the total ventricular activation time increases, and with that, the RV activation time increases and the LV activation time increases. We'll talk more about the patterns of RV and LV activation in patients who have left bundle branch block, but you see the timeline for activation. What you can see is that in each of these three patients, the latest activation shown in dark blue occurs in the lateral wall of the left ventricle in each of these three patients. Now, here is the problem with CRT therapy. CRT therapy is like taking a hammer, and you take this hammer and you hit everyone the same way. This is, to me, eye-opening data. Eye-opening data. This is data on 130 consecutive patients who had left bundle branch block. What you can see is most patients had transeptal conduction times greater than 30 milliseconds. This was all done in patients with ECGI imaging. Tremendous number of patients. But what you can see is there is at least a number of patients who either have normal or minimally impaired transeptal conduction time. Most patients have transeptal conduction time of at least 30 milliseconds. Then there are a whole lot of patients who have transeptal conduction times of 80 milliseconds or greater. What this says is that there are a number of patients who have structural abnormalities of the septum that result in slow conduction. This slow conduction through the septum is what's responsible for dyssynchrony or delayed activation between the right and left ventricle. That is what causes the QRS notching on the electric cardiogram that's characteristic of the left bundle branch block. But despite all these patients having left bundle branch block and most of these patients having QRS durations of 150 milliseconds or more, there is a lot of variability, a lot of variability in conduction in these patients. Here's an example of the struggle we have when we have patients who have right bundle branch block. These are two patients with right bundle branch block. This patient has right bundle branch block and you see they have rapid conduction of the left ventricle. Putting a CS lead in this patient with right bundle branch block, putting a lead in this patient's coronary sinus to resynchronize the left ventricle makes no sense at all. Their left ventricle is not desynchronized. There's no late activation of left ventricle at all. There's late activation of the right ventricle. Here we have another patient with right bundle branch block. Here the right bundle is far more fragmented or fractionated and wider. And here you can see in this patient the typical late activation of the lateral or the basal lateral wall of the left ventricle that we see in the majority of patients who have left bundle branch block. So this patient would not benefit from a CS lead or CRT. This patient would. And again, here's this computer model of CRT in the failing heart. Here we have left frontal branch block with LV delay 61, 20, 180. And you see these people respond dramatically to CRT therapy. Here's a patient with no RV LV delay, no delay between the RV and LV. It's not only is there no benefit from CRT, there's actually a decrease in stroke volume in this computer model. If the RV LV delay is now made to about 50 milliseconds, we see there's zero benefit, but not a detrimental effect. When the RV delay continues to increase, then we begin to see a benefit, a small benefit to a greater benefit, depending upon the amount of electrical delay between the right and left ventricle. So important concept to think about. Now, this and the subsequent slide come from papers done by Niraj Varma when he was at Case Western, working with the ECGI, which is basically a way to map the body surface that uses 256 electrodes on the patient's chest wall, a CT scan to allow an image of where the heart is relative to the chest wall. Now, this is all epicardial mapping, and this is all isochronal mapping. And the purpose of this is the LAD is here, therefore this is the right ventricle, this is the left ventricle. This is three different views, an anterior, a left lateral, and a poster view of the heart. And one can see in at least the first view, you can see where the LAD is. And the thick black lines in these pictures indicate areas of conduction, severe conduction delay, or a block. These isochronal maps look at a patient with a left bundle branch block. And what we see is, this is a typical left bundle branch block. Patient has left bundle branch block and native rhythm. E-synchronization means that the right ventricle and left ventricle electrical delay is huge. It's minus 113. That's how much the synchrony is. And you see typical basal lateral delay as you'd expect to see in people with left bundle branch block. You put a CRT device in, and all of a sudden, their delay has gone from negative to zero, means there's no delay to, and now there's a little bit of advance of the RV relative to the LV. But we've done a great job of re-synchronizing the heart in this patient with left bundle branch block. So simple, easy, straightforward. Well, here's the problem. Four different patients with left bundle branch block. Again, the isochronal maps. Y'all have left bundle branch block, but what you can see is in this first patient, the delay is from apical to the lateral base. You can see where the area of conduction slowing and actually block and delay is. So the electroactivation has to creep up here. This is the latest area. But in this other patient with left bundle branch block, the area delay is actually inferior to anterolateral. That's where the conduction block is. That's where the delay is. And here in this patient, we have combined activation now of the latest area from apical inferior and superior LV. And there's really only a very small area of delayed activation, a very delayed activation here. And in this other patient, in fact, you can see now the area of slowest activation, here's the line of conduction block, is somewhere halfway from the base to about two thirds of the way down towards the apex. And it's mostly inferior or posterior. So the take home message is, you'd want to put your lead to pace the heart in very different areas in these four patients. Here's a very localized area. Here's a more anterior area. Here's a more inferior posterior area. And here you wanna be closer to the base. So where you pace in patients with left bundle branch block, and these four patients with left bundle branch block, curestoration of 140 milliseconds, dyssynchrony, RVLV dyssynchrony of 73, curestoration 180, RVLV dyssynchrony of 113. Here your curestoration is 160, but only a very small degree of dyssynchrony. And here, another patient with curestoration of 180 milliseconds and substantially less dyssynchrony. So there are other things you need to be aware of. And that is, if you put your LV lead in SCAR or near SCAR, you're much less likely to have a responder. If you put your LV lead remote from the area of latest LV activation, you're much less likely to have a responder than if you put it adjacent to an area of late activation, or if you put it right there where there's late activation. This was actually a clinical trial, looked at patients with non ischemic cardiomyopathy, and then ischemic heart disease. Again, if the lead is remote from the SCAR, you have a better improvement, a bigger decrease in LV end systolic volume over time. If you put it adjacent to or within the SCAR, the decrease in LV end systolic volume was much smaller. And overall, non ischemic disease, because there doesn't tend to be as much SCAR, you have an even greater improvement in LV end systolic volume. And those patients do the best. Patients who have their LV lead remote from SCAR and patients who have non ischemic disease. What about, we talk a lot about QLV and where the LV is relatively. This is from some work we did in the SMART-AV trial that Michael Gold has been working on for over a decade. And if you put your LV lead in a late activated area, the LV, as measured from the onset of the QRS to the local LV activation. So if you look at where your CS lead goes and you measure from the onset of the QRS to that area where the CS lead is, if it's very long, the patient is, QLV is long, the patient is more likely to have a dramatic decrease in LV end systolic volume, a dramatic decrease in LV end diastolic volume, a better improvement in quality of life and a better improvement in injection fraction. And that correlated with improved neurocardioassociation class heart failure. And you can look at it by however you slice it up, whether you look at quartiles, but it is consistently been observed in multiple studies. In fact, one can go ahead and map the different branches of the LV. Here's an anterolateral vein where the local LV electrogram is 140 milliseconds after the EKG onset. Here's a poster vein where it's only 70. And so the best vein is this one where the QLV is the best. Now, this slide is a very complicated slide, but it is a great segue into where we're going. Hopefully I'll have time to talk about this. This is a, again, body surface map that looks at electrical activation. So this is a patient with left frontal branch block, a classic patient. You can see their LAD. You can see an anteroposterior view. And in the left lateral view, there is an area of very delayed activation shown in blue, slight purple in the lateral wall, left ventricle. So now you see a quadrupolar lead and those circles, each of those circles is one of the four electrodes. It's so happened that the Q to LV was about 80 milliseconds here, about 80 milliseconds here, about 80 milliseconds here, about 80 milliseconds there. And now we pace from different electrodes here from D1 to the ROV coil, from D1 to P4, and from M3 to the ROV coil. Here's D1, here's a P4, and here's M3. So these are each of the four electrodes pacing in a different spot. And even though the QRS to the LV lateral activation at each of these four electrodes is identical, this configuration here, okay, this configuration, I'm sorry, this configuration here, D1 to ROV coil, D1 to ROV coil has the least amount of blue, has the least amount of blue, blue being late activation. And that gives you the best biventricular synchronization using this distal LV electrode to the ROV coil. So these large variations are brought about despite identical or very similar QLVs. And what that means, this patient had non-ischemic cardiomyopathy, that the differences in activation, these different pictures, likely resulted from functional conduction barriers that varied according to the direction of depolarization, according to the direction of the wavefront of depolarization, right? Because if we had ischemic scars, you'd expect there to be fixed barriers. And you might not see the same type of picture. If there were fixed scar, then you'd expect the QLVs to be, if the QLVs were the same, you'd expect the electrical activation to be the same. But it's not, this is a non-ischemic cardiomyopathy, it implies that there are functional conduction barriers that cause that. This is a patient who has an IVCD. This patient had an IVCD, was referred for cardiac resynchronization therapy. Is this patient ever gonna, even though the cure-restoration is 150 milliseconds, is this patient gonna benefit from CRT therapy? Absolutely not. This patient has heterogeneous and abnormally slow activation of the right ventricle, with delayed activation in this mid-lateral area. LV activation is slow, and there's an incomplete enteralateral area of slow conduction. The latest area of activations in the lateral base is cure-restoration, is about 160 milliseconds. This is not really a patient who is gonna benefit from, this patient with this funny IVCD, without any of the characteristics of a left bundle branch block. Actually, not only is he not expected to improve, but did not improve with CRT therapy, it was a non-responder. And when you look at the electric activation, it's not surprising. Now, I show you this slide. Okay, I think it's important to understand that these are the areas of slow conduction of left bundle branch block. You've seen this before, it's lateral base of the heart. Look at what happens with RV apical pacing. The star here shows you where RV apical pacing is. Again, this shows you the LAD territory. What I want you to take home from this is that the RV pacing results in a very different RV activation compared to what's observed in patients who have intrinsic left bundle branch block. During RV apical pacing, RV activation is slower, and therefore less efficient than left bundle branch block. As a result, when we do by the pacing, we have less efficient RV activation compared to left bundle branch block. And there is a significant difference in activation, even of the left ventricle in these patients. Less later activation, less efficient, and that results in LV activation, prolonged LV total activation time, and different orientation of the areas of slow conduction. So, and the amount of dyssynchrony, intraventricular dyssynchrony, which is a major determinant of response to CRT, is higher with left bundle branch block intrinsic than with RV pacing. I wanna say a couple words about endocardial CRT. These are three different canine models of dyssynchrony, and they compare epicardial CRT to endocardial CRT. And what you see with endocardial CRT, regardless of pure electrical model, left bundle branch block, and ischemic heart disease, these are dog models with LAD and FARX, or left bundle branch block and heart failure, endocardial CRT is always, always, always two to three times better than epicardial CRT when you look at total LV activation time, but also one and a half to two times better in terms of LVDPDT. Now, this slide just says so much, and it says, this is a slide where they looked at DPDT max and paced in different venous branches of the coronary sinus. You can see how dramatically different it can be in patients when you look at the worst spot to the best spot. There are dramatic differences. Now, not every patient had a worst spot, but what you can see is some spots in the CS, patients can have a 30% decrease in DPDT, and then, or no, not much change, and then you go to a good spot and there's a 20% increase in DPDT. This is the power of CRT. You get anywhere from a 10 to 50% improvement in DRT. These are 33 patients, 21 non-ischemic, 12 ischemic, LVDPDT measured with a pressure wire, multiple CS pacing spots. LV endocardial pacing is better. I won't read this slide, but it's better, and it's better because you have physiologic activation from endo to epicardium, short a cure restoration, homogeneous repolarization, and this is one slide. The one thing I wanna talk about is dual site LV pacing, triple site LV pacing, quadruple site LV pacing, septuple site LV pacing. Don't worry, this is not human beings, this is a dog, but here's the take home message. If you find the best site in the coronary sinus to pace from, you don't need two leads in the coronary sinus, you don't need three leads in the coronary sinus, you don't need four leads in the coronary sinus, okay? Multi-site LV pacing only helps if you can't find a good site in the coronary sinus to pace, then you can pace from multiple sites. But look at this, endocardial pacing beats everything. So that's been shown in human beings as well. This is a beautiful study done by the Bordeaux group, shows virtually the same thing. Endocardial pacing is almost always better, but the best site is a whole lot better than the worst site, and the best site is typically not in the CS, it's almost always endocardial, whether you measure DPDT or pulse pressure. There is an investigational system for implanting endocardial leads. I won't go into great detail just because of time, but it involves going retrograde, crossing the aortic valve, putting this receiver here, it's quite involved. Here's a CRT device without a CRT lead. The patient has a battery here. The battery delivers electricity to a transmitter. Ultrasound energy is transmitted, and this implanted device takes the ultrasound energy, converts it to electrical energy. This transmitter needs to be placed in a good enough acoustic window so the ultrasound energy can be transmitted to this tiny implanted device, which is fixed to the LV endocardium, can convert the acoustic energy to electrical energy, and then this triggers a pacing spike right after RV apical pacing. So it's LV endocardial pacing triggered right after RV pacing. And you can see patients had a dramatic improvement in your cardio-sensation class, and about two-thirds of patients had a dramatic improvement in the LV ejection fraction. There is a way to do endocardial pacing by a superior transeptal, or you can do a transeptal inferiorly and take the lead across the septum using a snare. This was done in Europe. The problem, and this is great news, about 47% to 61% of non-responders, non-responders were made responders with endocardial pacing. The problem was that TIA incidence was almost 7%. Non-disabling strokes were 4%, and they were often related to subtherapeutic INRs. The INR has to be 2.5 or higher. This select LV shows that 80% of patients improved with endocardial pacing, 80%. So we take the number of responders from somewhere between 50 to 60, to maybe even 70%, to slightly over 80% with endocardial pacing. Now, I have only two or three minutes left. I want to just tell you, we know some patients don't do well with CRT, narrow QRSs, LV leads in the wrong spot, bad RV function, extensive scar, LV lead in scar, non-left bundle branch. This is a great way to think of non-responders. I call it the three Ps, poor patient selection, poor lead position, and poor programming. And you can go through that at your leisure when you watch this on video. This is from a review article by Jeannie Poole and Jack. Management of non-response to CRT goes through this. This is something every EP fellow should know backwards and forwards. Also a good area of test questions for the ABIM. And it's important to be able to think this through in every patient you put a CRT device in. For example, if a patient comes back and they're only pacing 85% of the time, by V pacing, why is that? Well, they have lots of PVCs. Well, then you want to suppress the PVCs or ablate the PVCs. The most important thing, the most important message I want you to take home about this is, this is a simple one, and that is make sure you have a good 12 VDKG before you put a CRT in. Make sure you have a good 12 VDKG of by V pacing because that gives you information about whether you actually have by V pacing. Maybe your LV lead is near a scar and you need to have LV activation earlier or need to increase LV pacing output to capture wider area. So those are all the things you need to think about and they are really important things. Suboptimal LV pacing, don't be fooled by pseudofusion. In pseudofusion in patients who have AFib, this is just a slide to remind me to say AV intervals can be important sometimes. This slide is another slide that looks at RV pacing and talks about how the effects of RV pacing on conduction can be confusing. This is the reference. I highly recommend you look through this because I think there are some important messages there, but I wanted to end with CRT optimization. Sometimes AV delay, sometimes VD delay can make a big difference. It is something you need to do in patients who are non-responders. As a strategy, however, it is not, randomized clinical trials have been disappointing. We did the SMART-AV trial and could show no benefit of echo-guided empirically programmed or device using the intracardiac electrogram. QWIC-OP uses intracardiac electrogram. The algorithm is programmable, but still not dynamic. Adaptive CRT is dynamic, non-inferior versus echo-by-V. Clinical trials still pending. SONAR hemodynamic sensor, the RESPOND trial, again, non-inferior SYNC-AV, which is a new Abbott algorithm. It is a dynamic algorithm. It's programmable. There's a large safety and efficacy trial that is starting soon or recently started, but an important trial. We're still waiting to hear about the final Medtronic trial. I want to end with this slide because this shows you body surface mapping and activation map for a typical left bundle branch block shown here, and this portion of the slide is VV-SYNC, and that's the difference between the mean RV activation and the mean LV activation time. This is the total global activation time of both ventricles, VV-TAT, and it's a measure of the total time required for, so VV-TAT is the total time required for both ventricles to activate. LV-TAT is the measure of the total time required for the LV to activate, and LV dispersion is the standard deviation of LV activation times, and it's a measure of the synchrony just within the L left ventricle. So here's some patient with intrinsic left bundle branch block. Here's the patient with RV pacing. They look, in this particular case, they look very much the same. Their RV-LV desynchrony is greater with left bundle branch block than with RV apical pacing, but the total ventricular activation time and the LV total activation times are not tremendously different. Here is conventional CRT, and here the VV synchrony, here you see the VV synchrony, the mean difference between the RV and LV activation time is tiny. The total ventricular activation time goes from 123 to 84. The LV total activation time is decreased, and LV dispersion is decreased. So that's conventional CRT with a CRS lead. Endocardial CRT, okay, with endocardial CRT shows now the VV synchrony is still small. The V to V total activation time is even shorter. It's gone from 84 to 53. The LV total activation time is dramatically reduced, and the dispersion of ventricular activation even less. Here is his bundle and LV and left bundle pacing. The V to V synchronization, now with his bundle pacing, a little bit more of the QRS duration is 101, here the QRS duration is 91, but you can see the VV synchrony now is still very, very, very low. With left bundle pacing, it's 10. The V to V total activation time is almost as short as it is with endocardial CRT. The LV total activation time is a little bit longer, 62, but less long than CRT. And LV dispersion is less. So what are the take-home messages? Avoid desynchronization in a narrow QRS complex, but don't forget to avoid AV desynchrony. That's not good. With these minimizing ventricular pacing long, peer intervals probably longer than 240 are not good. True resynchronization of a wide QRS complex, left bundle, they're the responders. That's what drives all the clinical trials are left bundle patients. And they're typically 70, 75% of clinical trials, much more so than right bundle. Some of those patients do have LV delay, probably no benefit for IVCD. Humility in CRT is critical. There's great biological variability. For resynchronization, re-coordination, avoid RV pre-excitation, aim at endo-epi activation, aim at apex base activation. LV alone, in most cases, probably as good as by V. For by V, generally narrower is better. Normalize a QRS complex. However, we haven't proven that a normal QRS complex equals no heart failure and living a long life. And with that, I'm gonna stop. Thank you. All right, thanks. That was amazing. Great job. Maybe I can ask just a couple of questions. One is, I guess, clinically and for trial design, what do you think the most important parameter to determine non-responses, what should we use? That's a great question. I think the answer is hemodynamic variables, LV end systolic, LV end diastolic volume, and some measure of heart failure hospitalization and total cardiac hospitalization. I think it's very reasonable to have a clinical primary endpoint and an echo primary endpoint, make them co-primary endpoints. So a clinical endpoint that's reasonable would be heart failure hospitalizations. I'm sorry, heart failure hospitalizations, total mortality, and total cardiac hospitalizations. LV end systolic volume is the one thing that can be reliably measured. And I think that has to be either co-primary endpoint or a second endpoint. The most important secondary endpoint by far is LV end systolic volume seems to predict cardiac mortality. Okay, and then I guess I would be remiss if I didn't ask this, but it's the same thing I asked Olu on Wednesday, which is what's your current approach to using physiologic pacing for resynchronization? How do you talk to patients about it? That's a great question. You would be remiss if I didn't answer that question. The answer to that question is, the answer to that question is, we now currently do, we do, for patients who have an EF below 35%, I think we only do conduction system pacing if CRT doesn't work out. Either there's some problem or a phrenic pacing or the spot is just not good physiologically, high thresholds, whatever. So it is not, we don't say, we don't tell patients we're gonna put a CRT system in you and then purposely put a left frontal branch block lead in place of a CRT lead. Now, do I have clinical equipoise to do that clinical trial? Yes, but I think that would involve informed consent and talking to a patient. For RV pacing in patients who have a normal EF, we, in 98% of patients, we put a left bundle lead in, 98% of patients and people have a normal EF. In patients on EF at 35 to 50%, we will either put a left bundle lead in and a CS lead in or a CS lead in or a left bundle lead in without any one of those being preferential. We'll often talk about what we wanna do ahead of time. So, but that's how we approach it.

cardiac pacing

cardiac resynchronization therapy

CRT

patient selection

lead positioning

programming

RV apical pacing

pacing-induced cardiomyopathy

non-RV apical sites

AV and VV timing optimization