[1] Thank you. I’m going to try to present the notion that selecting an ICD is more or less like selecting an antibiotic, has not only to do with the disease but also with patient characteristics like renal function for certain antibiotics and QT interval for another antibiotics. [2] So, the same thing goes for devices. These are my disclosures and I’m gonna be talking mostly about the major manufacturers but not because of my disclosures, but just because I use them more and I know them better. [3] So, I’m trying to present the notion that the devices are different. I’m gonna be talking about differences in sensing algorithms, difference in pacing algorithms, and how these relate to specific indications for certain devices and specific contraindications for other devices. [4] I will start with sensing and I will start by mentioning the difference between the sensing algorithm in pacemakers and in the ICDs. In pacemakers, you have a fixed sensitivity that you can program. But once you program the sensitivity, that will remain the same during time and this is because you only have to select sensitivity that is good enough so your pacemaker sees all the QRS complexes without sensing the T waves. Now, if this cannot work for the devices, the ICDs must be capable of seeing large events like the QRS complexes but they will also must be capable of seeing very small events like ventricular fibrillation and that is why the devices have what we called dynamic sensitivity which basically, they start from a nominal value and then as time goes by, the devices become more and more sensitive until they sense an event. Once they sense a ventricular event, the sensitivity will rise, meaning the device will become more sensitive, and then as time goes by, the sensitivity will increase. The device will start looking for smaller and smaller events. [5] So, this can cause trouble. One of the main causes for inappropriate shocks in patients with implanted device is T-wave oversensing and what is happening here is that, as the sensitivity increases, it reaches the point where it can sense a T wave and that is called T-wave oversensing. So, this leads to over-counting because the device is counting the QRS as one event and the T wave as a second event. So, you can imagine that if the patient goes into sinus tachycardia for about a hundred beats per minute and the device is counting the QRS and the T waves, it will end up counting ventricular events at a rate of 200 beats per minute and this will sooner or later lead to inappropriate shock which is one of the main causes of inappropriate shocks in cases with implanted ICDs. [6] Now, specific patients are, aren’t increased risk for this phenomenon of T-wave oversensing. For example, if you have a small R wave, then your chances of how your device sensing the T wave are obviously much larger and patients with right ventricular dysplasia sometimes have very large areas in the right ventricle where the R waves are very small and these patients eventually are at risk score of running into T-wave oversensing even if you don’t see this phenomenon during implantation. T-wave oversensing has also been described in patients with tall T waves, like it can be seen in the Brugada syndrome and also in the short QT syndrome. Some of the latest examples of T-wave oversensing that have been reported are from the short QT syndrome and paradoxically, you can run into the same problem in patients with long QT syndrome. These patients have small T waves but because of the long QT, they are farther apart, so the sensitivity increases to the point in time where it reaches the next T wave and leads to T-wave oversensing. [7] This is an example of a T-wave oversensing in a patient with long QT syndrome. As often it happens in patients with long QT syndrome, once they get the device, they forget about taking the beta-blockers. So, you can see that this patient is having sinus tachycardia with marked T wave alternans. So, he’s obviously not taking his beta-blockers. And because of the T wave alternans, he’s getting trouble counting and eventually gets inappropriate shock just because of T-wave oversensing. This is a close up look of what is happening and you can see that the device is sensing the QRS but also the large T wave and the next QRS is sensed and is counted as a ventricular fibrillation event. The next event then, the next T wave is not sensed and then the same sequence appears again because of the T wave alternans until eventually there is enough number of counted events and the patient gets a shock. [8] This is an example of T-wave oversensing in the short QT syndrome published seven years ago and this is the stored intracardiac recordings, where again, you can see that the device is counting the QRS and the T wave. This leads to double counting and eventually to inappropriate shock. [9] To overcome this problem, the devices, the companies have come with different algorithms. So, this is what is causing the T-wave oversensing. So, in the Medtronic, you can try to solve the problem just by raising the starting point of the sensitivity hoping that as the device becomes more sensitive, it will do that after the T wave overview occurred. The St. Jude has an algorithm that is called decay delay that will cause, the device will start increasing the sensitivity after a programmable period of time. So, hopefully, this will get you over the T wave. So, you should take this into consideration when you have a patient that is candidate for ICD that has some of the characteristics that I mentioned that increase the risk for T-wave oversensing. [10] Regarding pacing, as you know, for many years we were told that the basic pacing mode is the physiologic pacing of the DDD mode [11] and you are very aware, you’ve seen these slides before about the DAVID trial that showed that ventricular pacing in the right ventricle, whenever it’s not necessary, increases the risk for death or hospitalization for a heart failure. [12] And that is why we have changed our attitude. We do what, we do whatever we can in order to avoid pacing in the right ventricle when it’s not really necessary. Now, there are several options. You can put a VVI device and program it to pace very slow at the rate of VVI 35 beats per minute or you can use selective AAI mode at across the rate. This has the benefit of obviously preventing sinus node dysfunction that is sometimes triggered by the medications that we give for patients. This sinus node dysfunction does not have to be present at the time of the implantation but as the patient comes back, with apparent ventricular tachycardia and we start adding medications like amiodarone or increase the beta-blockers, they may develop symptomatic sinus bradycardia and if you implanted a VVI device, then you’re stuck with it. On the other hand, programming the device to AAI pacing will obviously be a problem with the patient if it develops complete heart block. So, one of the options is to use the DDI mode and select the longest AV delay, then you can get away with even AV delays of 300 beats per minute. This would, in theory, will allow you the best of all worlds because you can pace the atrium with very little pacing in the ventricle. [13] The problem with this pacing mode and with a very long AV delay is that, perhaps they can be proarrhythmic. When you think about the way we induce ventricular fibrillation in the laboratory by shocking on the T wave, we usually program the T wave shocks more or less around 300 milliseconds from the previous QRS. So, in theory, selecting the DDI mode with very long AV delays could be proarrhythmic and this is an example of one of our patients that I think, I think it’s impossible to prove it but I think that this is what happened to him. This is a patient, a post MI patient, that got his ICD for prophylactic reasons and that is why we were very surprised when he came back with the shock for polymorphic ventricular tachycardia because one would expect him to come with more than one PVT. Now, if we look closure of what’s happening, what is happening here is that he is pacing with a very long AV delay. So, here he is pacing in the atrium and he is sensing in the ventricle. Again, pacing in the atrium and sensing at the ventricle, here he has a spontaneous ventricular extrasystole that just by chance, happens to be at the time when the atrium is pacing. So, this ventricular sense event is refractory. The PVC is not seen by the device and therefore, the atrial pacing is followed by ventricular pacing after 300 milliseconds which probably occur on the T wave on this PVC. Nothing happened here but the same sequence of events occurs again, atrial pacing followed by ventricular pacing. Then again, a ventricular extrasystole that just by accident falls at the same time of the atrial paced event and therefore, the ventricular extrasystole is refractory. It is not seen by the device and therefore, this atrium pacing is followed by ventricular pacing that it probably falls on the T wave and induces polymorphic ventricular tachycardia or VF. This patient had two events that would be very similar and they were both started by ventricular pacing and probably we think this is the proarrhythmic effect of pacing with a very, very long AV delay and this is probably a rare event but one needs to take into consideration that this may happen. [14] Therefore, to solve this problem, there are several algorithms like this one from Medtronic that is called Managed Ventricular Pacing. So, basically what it does, it ignores the AV delay as an AVI mode, meaning, it paces only the atrium but it senses the ventricle. So, you’re pacing in the AAI mode but is, you’re watching what is happening in the ventricle and if there is one atrial sensed event or paced event that is not followed by a ventricular sensed event, then the device will pace the ventricle with a very short AV delay and if this happens again, if there is another atrial paced event or sensed event that is not followed by a spontaneous ventricular sensed event, then the device will switch from the AAI pacing mode to the DDI pacing mode first with a very short AV delay to avoid pacing on the T wave and then with the programmable AV delay. So, this is really the best of all worlds. [15] And after a few, after a programmable period of time, the device will search for spontaneous QRS complexes by again changing into the AAI pacing mode for one cycle and if there is a sensed event, it will again remain in the AAI pacing mode. [16] So this is an example from one of our patients that has an implanted fibrillator. He has a relatively long PR interval. He has a wide QRS but he does not have heart failure and therefore, he does not really need a CRT device. He is in functional Class I. Now, here programming any device with a maximum AV delay with 300 milliseconds would still pace the ventricle because his basic PR interval is very long. He needs atrial pacing because he has symptomatic sinus tachycardia so he gets this device with Managed Ventricular Pacing. To make it easier, I will put the markers so the red line is atrial pace and the blue line is ventricular pace. So you can see that he is pacing in the atrium and conducting, he is pacing in the atrium at the AAI mode. He has one to one conduction. Here, he has paroxysmal AV blocks, so he’s pacing the atrium. There is atrial capture clearly seen here. There is AV block. So, the next event is atrial pacing followed by ventricular pacing with a very short AV delay which gives you this paced QRS complex. Again, he remains in atrial pacing with one to one sensing until he has here AV block again. It’s probably... and whenever he has AV block, he has pacing with a short AV delay. So, really most of the time, he is sensing in the ventricle. This is just magnified view to show you there is paroxysmal AV block followed by atrial and ventricular pacing with a short AV delay. [17] Now, one of the important things to take into consideration is the availability of post-shock pacing. By post-shock pacing, I mean, the probability of pacing at a different rate after a shock, the ability to pace with a rate that is different from the basic pacing rate. So, this allows you to pace slowly most of the time. But at the same time will allow you to pace fast after you get an appropriate ICD shock. Now, this feature is available in devices but there are several important differences. For example, the Guidant devices will allow you to pace fast after a shock for up to 60 minutes. The St. Jude devices will allow you to pace fast for up to ten minutes and the Medtronic devices, this was not available until recently just from the EnTrust model and newer models will allow you to pace faster after the shock. [18] Now, why is this important? This is important because patients may develop this particular arrhythmia, torsades de pointes from long QT and one of the characteristics of this arrhythmia is that it tends to recur after pauses. So, as you well know, in torsades, it always starts after its sequence of short long intervals. PVC will trigger an extrasystolic pause so the next thing is we’ll have a huge QT interval for which torsades will begin. So, whenever a patient is at risk for developing torsades de pointes, you should be able, you should have the ability of pacing faster after a shock because otherwise, he is at risk of developing arrhythmic storms of torsades. So, this is a consideration that should be taken obviously in patients who have congenital long QT syndrome but also in patients that are receiving or receiving future medications that prolong the QT interval like sotalol or quinidine or even amiodarone because you never know who will develop a long QT syndrome from these medications. [19] Now, the problem is that, sometimes you cannot tell in advance who will develop this complication and this is a case of a 34-year-old male that was transferred to us after out-of-hospital resuscitation for a ventricular fibrillation. He has a dilated cardiomyopathy with a moderate left ventricular dysfunction. His left ventricular ejection fraction is 40%. He was treated with beta-blockers and ACE-inhibitors and also amiodarone. This is the trace that was recorded while he was transferred, while he was being treated with ACE-inhibitors, beta-blockers, and amiodarone. So, you can see, he has some poor progression of the R waves, some mild intraventricular conduction delay but really nothing particular or the QT certainly does not appear to be very prolonged especially considering he was receiving amiodarone at that time. So, we end up implanting an ICD without post-pacing, without the ability of pacing faster if there is shock. He got a Medtronic device and he had a single-chamber device because he has no anticipated need for cardiac pacing. He has only moderate left ventricular dysfunction and his arrhythmia was ventricular fibrillation so he got a single-lead device, the VVI device and he was found dead at home three months later. Now, we performed a post-mortem interrogation at the device. [20] We were very surprised to see what actually happened. What actually happened is that, he developed this atrial tachyarrhythmia probably an atrial flutter with very high degree of heart block, probably complete heart block. So, he is paced. He is being paced at the program in lower rate limit which was 40 beats per minute. So, he’s in complete heart block. He is pacing at the ventricle with capture and then he starts having ventricular extrasystoles and he develops polymorphic ventricular tachycardia. He gets an appropriate shock that terminates the arrhythmia but he remains at a very slow rate and he develops polymorphic ventricular tachycardia again. So, this is just probably torsades from, caused by the very low pacing rate. The torsades sensed again by the device gets another shock but because he remains bradycardic, he goes into torsades again. So, what actually happened is that he developed an arrhythmic storm of torsades that we did not anticipate. We did not anticipate a complete AV block and we did anticipate that he would develop a long QT syndrome during bradycardia and that is why now [21] we always program a certain amount of post-pacing after a shock at least for a few minute. This is another view of what [22] happened. The [23] sequence of events until eventually [24] Skipped slide. [25] Skipped slide. [26] eventually after many, many shocks, the device did not sense the ongoing ventricular fibrillation probably just because this is by now an abnormal rhythm and the patient died in refractory ventricular fibrillation. [27] So to conclude, I will say that not all patients are equal and not all ICDs are equal and there are points to take into consideration for patients who are at risk for T-wave oversensing because of long QT syndrome, short QT syndrome, or Brugada syndrome, or RV dysplasia. We should probably use devices that have algorithms that will prevent this T-wave oversensing like the decay delay. For patients with sinus bradycardia, the algorithms that allowed AAI pacing while still sensing the ventricle are probably the way to go and we, I recommend to always use post-shock pacing at least some degree and I think this is the most for patients who are at risk for pause-dependent ventricular arrhythmias. [28] And thank you for your attention. [29]

Sami Viskin 151AB What Type of ICD to Implant: How to Select the Right Device Every Time Selecting the Right Device for the Right Patient: What Type of ICD to Implant