Video Transcript:

I will start with the scenario that we see now and then. So, this is a patient, 68 years old female patient with aortic stenosis. Body surface area is 1.5. Good LV function. No comorbid factors. We have consented her, she is in an operating room. The plan is to have a stented biological valve, pericardial valve. The surgeon opens her up. The biggest valve size he can fit in is a 19 mm. He looks at the chart that comes in the insert of this valve and that would put her in so called severe mismatch with the indexed orifice area of 0.73 in the red. So, you are a surgeon or you are the cardiologist who referred this patient, what would you expect as a cardiologist and what would you do as a surgeon. So, would you proceed with a standard 19-mm stented valve, pericardial valve, would you put a 19 mm mechanical valve expecting better hemodynamics, would you expect root enlargement and a 23-mm stented valve pericardial, the subcoronary stentless valve 21 mm, or would you replace the aortic roots with a 23-mm porcine valve? So, half would enlarge the roots and put a 23-mm stented valve and about 20% would use a stentless valve either as a root or subcoronary. So, I will follow on with that question and ask why you gave the answer. What is the reason that you gave best qualifies the answer. You believe a 19-mm valve was inconsequential in most patients, you believe it's associated with reduced short and long term survival, you believe that industry charts indicate the patient has severe patient prothesis mismatch, PPM. You believe it is best to avoid 19 mm valves but we have no reproducible ways of doing so or you believe stentless valves or roots have better long-term outcome. So what is the reason you gave the answer you gave; it is inconsequential, it has reduced survival, the industry charts say it is severe, you do no have good ways of fixing it, or the stentless valves are best. Let's see the answer. So, most people believe the stentless valves or root enlargement have better long-term outcomes. A lot of people go by the charts and only 14% believe it is inconsequential.

So, I have got a task here and I am going to be very controversial because I am going to try and convince you, I will try to convince you and we will see what you get convinced about. So, I am going to go through several myths about patient-prosthesis mismatch. I am going to deliberately challenge the data. The first is that it is a clearly defined entity. So, where does this all come from. This comes from Rahimtoola a seminal paper 31 years ago and what did he say. He said mismatch can be considered present when the effective valve area of a prosthesis is less than that of a normal human valve. That is what Rahimtoola said. Although, he did say it is often of no clinical significance which is interesting, but then he goes on to confound himself saying while the area of almost all types of prosthesis is less than that of a normal human valve. So does that mean that patient prosthesis mismatch is a necessary part of prosthetic valve replacement. He went on to have a nice graph like this and he indexed the valve area to the body surface area and defined an area of mismatch. So, that introduced the concept which we have carried until today which is indexing body surface area but where does it come from. I mean if you have a sort of BSA of 1.8 as a nice young patient and then you put on a weight you go up to 110 kg your body surface goes up, does it mean your valve now becomes stenotic, the same valve. I mean is it mathematical invention or is it a true phenomenon. Indeed if you look at the theory this is a lot of that comes from Pibarot's group and if you look at the theory for the same valve orifice here, as the body surface area goes up the gradient should go up but what happens in practice, so the valve area is going down, the gradient is all over the place. So, a lot of patients even with an indexed valve area of 0.7, 0.6 have little gradient across the valve. I mean the only thing that is true is that you do not get mismatch with a big valve. That is certain. So, if your indexed orifice area is 1.6 you do not get mismatch. I mean what is quite strange they have patients with indexed areas of 2.6, 3.0, I mean the only way you could do that is if you pull like an elephant's valve into a child well otherwise how could you possibly have an indexed area of 2.6 in an adult. This is amazing. But anyway what if you look here for the patients with a mismatch less than 0.85 indexed, the gradient could be anyway from very little to very much with all types of valves and that is the reality. The third myth I would like to point out is that we as surgeons cannot predict the likelihood of PPM. So, we use these charts but the reality is that the charts have nothing to do with what happens. The gradients are all over the place; whatever size valve we pick we see that unless you are going to put 25 valves in everyone you will see this dispersion. The fourth myth is that we get higher mortality with patient-prosthesis mismatch both early and late. You will see a lot of data in the literature that compares mortality with mismatch and without mismatch and as always you will see this was with mismatch. So, if we go back to Rahimtoola because he is the one that created this 30 years ago and now it has become an animal of his own and what does he say. He says to correctly address the issue, prosthetic valve area should be calculated from echo or Doppler studies six or twelve months after implantation, not in the operating room, not with these kinds of charts. He is saying that you have to do echo and workouts on echo. Look at all these studies, all of them are using the manufacturer charts, none of them have to find a mismatch based on real hemodynamics, real gradients, and real valve area. So, how then can they define the outcome and the reality is as we have seen you cannot use valve size to predict the postoperative orifice area. The correlation, as you can see, is very weak. So, we if we look again at the data this is data from Angelini's group published a few months ago and they looked at all the 19-mm valves. Based on the manufacturer's data, the orifice area should be 0.6 but actually measured on stress echo what the true area was and the average area was 0.8. So, what the manufacturer's say and the reality are not necessarily the same. So, it is true that you would rarely see patient-prosthesis mismatch a lot sizes that are concieded but the size of the valve alone does not explain mismatch. Many patients who you expect to have mismatch based on the industry data will not if you measure it in the clinic with echo and we do not know how to predict those patients who do have small valves that will transform to clinically relevant patient-prosthesis mismatch. So, the data that supports the adverse effect of patient-prosthesis mismatch are weak in the long term. So, let us look at some of the data. This is data from Dr. David's group published about nine years ago. So, although it is titled "Effect Survival", if you look at the 12-year survival whether you had mismatch or no mismatch, it was the same. There was a higher instance of cardiac related deaths in the mismatch group but overall survival was the same and when on multivariant analyses valve area was not a predictor of long-term survival; this is 12 years out and this is substantial, over 2000 patients. Although, it is important to know that about 30% of patients were dropped from the study because they did not have data on body surface area so it could be that with those patients the results could be different but it is difficult to see how a patient cannot have a body surface area given that every patient has a height and every patient has a weight, so how that data can be missing is a bit puzzling but anyway that's the data. This is a more recent paper published from Vienna and they found the same thing. No difference in long-term survival, so the axis starts at 65, probably a tendency towards high rate of mortality but the curves stay the same. The fifth myth is that as surgeons we can actually eliminate patient-prosthesis mismatch, so we can design better valves, stentless valves, we can sew the valve in the annulus, outside the annulus, enlarge the roots and then we can do this without risk but let us first look at what happens in practice. These are recent papers now. So, this was just published a few months ago in the American Journal of Cardiology. This is from Brisbane. They have 156 patients and 58% by definition had patient-prosthesis mismatch. This is the real world. In Vienna, the same thing, 54% of patients had patient-prosthesis mismatch. So, we have been talking about this for the last 20 years, but in practice as surgeons we are still putting in these small valves and indeed if you go through where the source started with Dr. David's group who over the last 30 years has led the field in terms of ways we can prevent this. Look at their data and they say that almost 30% of patients received valves with a ratio of less than 0.9. So, even with all attempts to fix the problem in Dr. David's hands there was still about 30% instance of mismatch. So, even with the techniques we have available it is not certain that we can actually fix the problem in all patients. We can certainly fix it in some but not all. So, let us assume for argument sake that we can reduce the incidents of patient-prosthesis mismatch, so the stentless valves work, the root enlargements work, is there a cost to this? And if we go back to Dr. David's data, I'll just concentrate on this period 1995 to 2000. The mortality for root enlargement was 7.5% and this is in 360 patients, so this is isn't a learning curve. This is a very experienced team who have led the field in this specialty and in the recent time had a 7% mortality. Now the mortality is down but if this is what David's group achieves what would I achieve if I do this only once every three years. So, for the patient I presented to you I have never done a root enlargement in maybe three years and I do that today, would I expect to be even in this 7.5% for a surgeon that has done 360 or will I be much higher? Chances are is that most of us will in reality be much higher. I mean some people do report even better results and you can see here the 30-day mortality if you have a root enlargement is 1%, 0.9%, which is even better than Dr. David's but is even better than a valve replacement, you can see in valve there is 4%. So, we might as well replace everyone's root and we will lower the mortality so you can choose what you want to believe. So, the other alternative is a freestyle valve. So, a lot of patients when were driven by the bandwagon of patient-prosthesis mismatch in the 1990s and received freestyle valves and it was said to be the best thing. It was even better than root enlargement. So, this is from Quebec again, Pibarot's group and 419 freestyle valves and of those 68 had the size 19 or 21. The operative mortality was 6% in this group. The five-year survival if you had a 19 was 58% and 82% for 21, which isn't very good and indeed the other observation is, that we all think a root replacement solves the problem and if you look; these patients had a freestyle valve, some are subcoronary and some are roots and if those that had a 19 still had a 100% instance of patient-prosthesis mismatch, 21 had a 70% instance of patient-prosthesis mismatch and this is real mismatch measured with echo. So this is not manufacturer areas, these are Doppler measured orifice areas and if you had a 23 freestyle there was almost a 50% incidence of patient-prosthesis mismatch. So, this led the office to conclude that an AVL with a freestyle subcoronaries associated with a high instance of patient-prosthesis mismatch and this is interesting to put this in perspective under high mortality because it then means that a lot of patients in the 1990s actually suffered because of our belief in patient-prosthesis mismatch and I believe that we had a solution, i.e. freestyle valve replacement.

This is another paper about stentless valves, now this is a randomized trial. So, stentless valves are supposed to do the trick and in a randomized trial they found no difference in the orifice area in patients randomized to either Perimount valve or a Toronto stentless valve. So, the sixth myth I would like to dispel is that these strategies to eliminate patient-prosthesis mismatch result in better long-term outcomes. There are really no good comparative data and even if there are late gains there might be offset by the higher early mortality that I have alluded to, but I think more important to consider is that patient-prosthesis mismatch might actually just be a confounding factor and the worst outcomes may have nothing to do with the valve area but to do with whatever factors make that patient have patient-prosthesis mismatch in the first place. This is a paper from the Vienna group again which I've shown earlier and they have looked at patients with mismatch and those without them and compared various characteristics. It does not matter what the details are but you can see a lot of P-values, a lot of differences; their older like seven years older than the people with mismatch there are more females who have a higher euro score and indeed on multivariant analyses the only thing that has predicted the outcome in these patients was not prosthesis mismatch but the euroscore and diabetes. So, a lot of these patients that we say do badly because the valves we put in are too small may be they were going to do badly anyway because of other factors within the patient and they said that the result shows that patient-prosthesis mismatch is stronger, I should say that with unfavorable preoperative conditions. So those unfavorable conditions probably put them at disadvantage at the time of surgery. So, the last myth I would like to dispel is the one of low-grade intra-aortic stenosis. So, we say that if the gradient is low there is no EF. That is the one that you really must, must, must avoid patient-prosthesis mismatch and indeed there are data to show that. So, here you have data from Ruel and colleagues and you find that if you have an impaired ventricle and patient-prosthesis mismatch your survival is worse and it is good to read the papers because when the study is not looking for what you are looking for you find other things that they were looking for. So, if you look here at the top of the graph you see that patients with normal LV with no mismatch and patient with impaired LV with no mismatch seem to have the same survival and that does not make sense because we all know that impaired LV function is a predictor of survival and obviously if they were writing a paper about impaired LV they would have not published this but anyway, what's important here is that they defined a poor LV function as less than 50% which probably isn't what most of us in this room and certainly not what Dr. Carabello would regard as low EF aortic stenosis. So, really if you torture the numbers they will tell you whatever you want to tell them, they will confess to anything and that is what statistics do. So, many of you would have seen this picture before. This is what tells you that if you have a low EF you must not mismatch the patients because if you have a low EF here and severe mismatch you are 70 times more likely to die than anyone else which is interesting because how is it possible anything in biology that your odds ratio is 80, it is not possible. So fine, what it reveals is interesting but there is something that is hidden which has to be vital and the answer is if you read between the lines. So, as I said you torture the numbers they will tell you anything. So, you have 1000 patients. This was published in Circulation. So, Circulation, the editors like the 1000 patients but they did not realize what was concealed by the bikini which is that in this group on which they base their entire study there was three patients. So, there were three patients with poor LV function and mismatch. Of course, they all died and they wrote a paper in which they say that 1000 patients published in Circulation, the best journal in our specialty, to say that you should not mismatch patients. So, I will tell you what the facts are that I see. The first is that the definition is unclear. The second is that most cases are of no clinical significance; I agree some are but most are not. The problem is we cannot predict those cases like the case we saw this morning, I think it was Lisa that presented the echo this morning. So, sometimes it is significant but the valve area is a poor surrogate and indeed there may be confounding factors rather than the valve area that explained this and those are approaches that we favored like 50% in the audience wanted to enlarge the root, 20% wanted to do stentless valves. They do not yield consistent results even in the hands of the experts, so I would implore that the cases should be considered on a case to case basis and we still need research to be able to predict who actually develops clinically relevant patient-prosthesis mismatch and the fiction is really everything else.

Thanks for your audience.