Episode 51: CONVINCE Study in the Context of Existing Evidence on Hemodiafiltration with Dr. Giovanni Strippoli

Episode Transcript:

Frank Maddux: Welcome to Global Medical Office Dialogues. Hemodiafiltration, also known as HDF, removes larger amounts of solutes and protein -bound compounds than conventional high -flux hemodialysis. There are indications that the people requiring dialysis who receive the highest convection volumes from HDF experience improved survival rates compared with traditional hemodialysis. 

Dr. Giovanni Strippoli, professor of nephrology at the University of Bari, Italy, discusses how the Convince study results will contribute to reaching a higher level of certainty regarding the effectiveness of HDF versus HD in people with end -stage kidney disease. Welcome, Giovanni.

Giovanni Strippoli: Thank you.

Frank Maddux: Tell us about the CONVINCE trial and what the study results actually were.

Giovanni Strippoli: The CONVINCE trial is a trial that has been designed to come to a conclusive statement about the benefits and harms of hemodiafiltration over and above standard high flux hemodialysis. And what the trial actually showed is that high volume HDF significantly reduced the risk of all -cause death, did not affect the risk of cardiovascular death, and had an effect on the reduction of infectious -related deaths and particularly COVID -related deaths. In addition, it also showed significant improvements in patient-reported outcomes, actually for the first time in a large-scale government-funded pragmatic trial.

Frank Maddux: What do you think the importance of this trial versus other randomized controlled trials have been as we look at the full body of evidence for HDF.

Giovanni Strippoli: Essentially the reason why we did this trial is that we already had quite a few trials, and their results were pretty strong in favor of HDF, but the medical community still didn't consider them enough or sufficient to come to a conclusive statement. So, what we really wanted to do was a properly powered, large-scale, pragmatic trial funded by a governmental body, the European Union. That would finally allow us to make a conclusive statement. And effectively what has happened is that the trial has shown actually, for the first time quite a strong reduction in the risk of all -cause death. 

But in reality, my take on this is that we should always look at the trial in the context of all other trials which were already existing and whether or not this new evidence that has been generated is game-changing compared to what we knew before. Because what we knew before was that hemodiafiltration reduced the risk of cardiovascular death. We knew this from a large series of trials, pretty large trial, relatively well-designed trials. But with this new trial now, we really wanted to make a final statement, and we really have to look at it in the context of that evidence, otherwise we will be dragged away from what it really is showing because it doesn't show a reduction in cardiovascular death, it shows a reduction in all-cause mortality. 

So, if we look at the trial per se, we might get more confused while the real interpretation from an epidemiological standpoint is that in the context of the totality of evidence, this trial really makes a convincing and final statement about the benefits of HDF over and above standard practice which is high flux hemodialysis.

Frank Maddux: Speak to us for a minute a little bit about some of the other aspects of the trial. Not only did it look at mortality as a primary endpoint, but it also significantly evolved the patient reported outcome measures as well as health economic analysis of HDF versus HD. Can you speak to those other elements as well?

Giovanni Strippoli: The European community was specifically looking at something that had to do with innovation and the true innovation we thought we would put into the request was actually a thorough measurement of patient reported outcomes. And the reason why we did this is that this is really the latest trend in research. When all the trials were conceived, they were very classical trials, where typically doctors are interested in seeing whether or not an intervention affects all-cause mortality in people with end -stage kidney disease. 

But over the past two decades, a lot of movements, including our own Song initiative have actually looked at patient perspectives about what matters to them of the care we give, and we deliver. And one thing that has really come out is that it is extremely important to them that they have an improvement in the outcomes that they perceive are relevant on top or as besides the effect on survival, all-cause mortality, cardiovascular events, etc. 

So, we put that into the application, and it was really the first time because there have been other studies in this setting that have looked at this but in a less structured way with lower numbers, with methodology that was still being developed while today we have a core outcome set developed by the SONG organization and other organizations. So, we put these outcomes into our measurement. We did a far-off analysis of patient reported outcomes. 

And what is really interesting, and this data is soon to be released in the public domain, is that besides the effect on survival, the intervention actually does affect patient reported outcomes, a large series of them. And the data will become public soon, but there is an improvement. There are two key observations and that one of them is that patients who are on hemodialysis for renal replacement therapy generally display a deterioration of a number of domains which have to do with cognition, well-being, et cetera. And so, this also happened in our trial. 

But in the HDF arm, specifically the high -volume HDF arm of the CONVINCE trial, these outcomes were significantly better compared to those in the high flux hemodialysis arm. So, this is highly relevant to the patient, highly innovative in terms of the methodology, and is certainly one of the key elements of this trial, and certainly one of the key reasons why the European Union did fund the trial to have a statement. This is also gonna be very useful because what it will do is it will inform, our ability to then do guidelines in the setting of interventions for end -stage kidney disease focusing on dialysis techniques. 

So hemodiafiltration, high flux hemodialysis, low flux has been completely abandoned these days, but there is now room and equipoise for formulating guidelines because we will be releasing patient reported outcome data and also a health economics assessment, which is basically all that is needed to then move on to formulating guidelines recommendations.

Frank Maddux: In your opinion, do you think HDF should be the standard of care that at least is offered in a trial form for most patients? Or how do you look at the proportion of patients that might benefit given that there are positive patient reported outcomes, there's positive quantitative heart outcomes on mortality, and I'm curious your perspective on the standard of care.

Giovanni Strippoli: I will give you a straight answer, but then you will allow me to elaborate on the reasons. The straight answer is that I definitely think that there should be straightforward recommendations from guideline agency that HDF would be standard of care. Okay? Now, of course, by the time such recommendations are released to the time they are actually implemented, this is going to take a large number of years and it's also going to have a geographical variability because there are countries where the technique is not used at all and countries where it's more broadly used. 

So, I think having a clear recommendation is going to result into starting to adopt the technique more broadly and it's going to take time until it actually ramps up to broader use. And I will also say that this is one of the reasons why also the trial was conceived. I will share with you a personal experience. When I was working, I worked for several years in the dialysis setting. And one thing I found that was really incredible is that colleagues that referred to me as a senior medical leader would contact me and tell me, how is it possible that in this region in Italy, I am not allowed to do HDF to prescribe it because it's not reimbursed while in this other region, they do 20%?

And this actually happened in Italy, and it happened on a global scale. I remember people calling me from countries where, and you're in the United States. What a better example? Where the technique is actually not used at all. And people are constantly emailing me or calling me and asking me how is it possible. I always replied that the only way we had to make a change in that was to really summarize evidence in a conclusive way such that there would be no further possibility to object, and I wasn't really sure until now.

However, now, and this is where I want to explain a little bit why I would strongly recommend the technique. Now we are no longer in a situation where we have few or conflicting or insufficient data. We have a very large number of trials available at the moment. We're talking about more than 40 trials. We're talking about 4 ,000 patients overall. And the evidence altogether is extremely strong. And what is happening now is that people are debating whether or not guidelines should be released because they are focusing on the single trial. 

This is a typical problem with the nephrology community. I have outlined this problem multiple times in my academic career. This happened with erythropoietins where people at the Besarab trial came out, it showed significant harm of high hemoglobin levels and people wanted more trials and so they did the CHOIR, the CREATE, the TREAT and more and more and more because people were always focusing on the single trial. This is happening also with HDF. 

Now people are focusing on Convince and the fact that it did not show a reduction on cardiovascular mortality and so people are lost and conflicted because this is in contrast with what they knew. But when you actually look at the totality of evidence, the evidence is incredibly strong. There's 4,000 patients. There is a large number of trials. There is zero heterogeneity. When you put the Convince trial together with the other trials that have been done, looking at cardiovascular mortality, there is a significant reduction in the risk even if Convince didn't show it. Overall, the data show it. And in this analysis, I recently did a contemporary meta-analysis on the topic, there is 0 % heterogeneity. What this means is that nothing is going to change that. If we get the new trial, and we're going to have a new trial from the UK, that trial is not going to change the findings for cardiovascular death. There is a reduction.

0% heterogeneity, unchangeable, unless you find, unless you really, really have another trial that shows that the HDF is dramatically increasing the risk of cardiovascular death. Similarly, for all -cause mortality, we did not have evidence of a significant reduction in the risk. Now we have it in CONVINCE, which is high -volume HDF delivered, in patients primarily recruited in private for -profit clinics where it is well known that the care delivered is very, very careful. There is a lot of focus on really delivering excellent care because this retains the patients for a very long period of time in the private for -profit clinics. So private for -profit clinics still adopt programs like the fistula first program that have actually been almost abandoned in standard practice. But in these centers, most of the people have fistulas, they are kept relatively well, high volumes are delivered. 

So yes, the Convince trial shows a reduction of all -cause death in people who are otherwise relatively well and who are receiving high volumes. The totality of evidence doesn't. And it's only in that population that there's gonna be that benefit. But the data overall is so solid for cardiovascular death and so strong for all-cause death in selected population that there should be strong guideline recommendations. I personally advocate for that. However, as I said, we need to have the health economics data, and the patient reported outcome data in order to be able to trigger that process.

Frank Maddux: Talk to me for just a moment about a topic you and I have discussed before, and that is the convection volume and the dose dependency. And to some degree, we probably see benefit, especially when we begin looking at larger BMI patients and people that might be contrary to the value that Convince brings in the US in looking at that 23 liters is probably not a hard and fast number, but it is recognized that there is a dose -dependent relationship on convection volume. Describe your views on that.

Giovanni Strippoli: When the CONVINCE study was designed and we submitted it to the European Union, I was not particularly loved by the steering committee members because I was always quite strong on some epidemiological statements. And one of them was that I did not want the concept of volume to be included in the randomization scheme because there is a principle of epidemiology, which is called Empiric Evidence of Effect Modification. It says that Empiric Evidence of Effect Modification is rare. What this means is that, in simple terms, if a drug has an effect in a population, it will have an effect on every other population. Take the example of aspirin. If aspirin reduces body temperature in people like me, it will also reduce it in people with diabetes. It will also reduce it in elderly or younger people. And it is very rare that this will not be true. Very, very rare. There are cases, but these are very rare cases. I will give you an example. An example is statins. Statins work for everybody. For me, for you, for dialysis people, no. And this is very rare.

So, we have shown, I mean, I would give a statin to anybody, even independent of the baseline cholesterol level. It's highly, highly, highly effective for cardiovascular protection, but in hemodialysis patient, it doesn't work. This is extremely rare. It almost never happens. It's extremely, extremely rare. So, the point I made was, why do you want to put this volume concept in If hemodialysis reduces cardiovascular death, it will reduce it in anybody. It will reduce it in people who can achieve high volumes because they have a good vascular access, et cetera, but it will also reduce it in people who do not. And that's partly addressing also what you were asking me previously in terms of should we give straightforward recommendation and say, give it to anybody or should we only say, give it only to people who can achieve high volumes.   We should say give it to anybody because if it's effective, it's effective in anybody. Then in practice, we may be able to give it to people in the beginning, we may ramp up the percentage by giving it to people who can take it more easily at higher volumes, et cetera. 

So, to come to a more direct answer to your question. What Convince showed is that there is a survival advantage of HDF in people who can achieve high volume, but that would put us off in a way because one of the standard criticisms that is coming out and there are more which I think we could also touch upon but one of the key criticism is well, yes, but it is only high volume and we can't really do it in anybody.  It's difficult. They did it in the private clinics because all their patients have a fistula and they could do it, but it's unreal. This is actually incorrect because what we showed by pooling the data of both CONVINCE and all the other trials that have compared HDF versus HD and that have all-cause mortality as a primary outcome. 

So, we identified five of these trials and we did an individual patient data meta-analysis where we had the individual data of each and every patient in each and every trial. And so, we also had the volume for each patient. Now in CONVINCE, we aimed to achieve a high volume, but in the other trials it was not necessarily like that. And so, when we put those data together, what we found is that there is actually a trend. So, the higher the volume, the higher the benefit. Now this brings me back to my original opinion. And I know that, of course, volumes are important, but they are gonna generate a lot of discussions because people are gonna say, well, but what if I cannot achieve a high volume? What if a person is very small in size, or small weight, or body mass index, et cetera. 

Now we're gonna do more studies to look into this, but the principle to me is that the intervention reduces cardiovascular death. If you can achieve high volumes, you can also reduce all-cause death. The higher the volume, the better. But in essence, the intervention is good, and it's good overall. So, by the time we give it to 100% of patients, we'd better start giving it, and then we will, of course, fine tune. It's a little bit like the KT/V concept. Firstly, you've got to give them dialysis because otherwise they're going to die. But then you're going to refine and say, okay, let's try and give them better adequacy, better anemia management, et cetera. Always with intelligence because it's not necessarily always said that the more the better. I'm not a believer in that concept, but certainly there is an association between higher volumes and improved outcomes. And this is going to be published soon as well.

And hopefully we'll put the question to the debate, which is one of many. I mean, people are debating whether it's effective in people with diabetes, people are debating whether it's only effective in people with fistulas, et cetera. And these are all standard criticisms, and I'm happy to elaborate on that if you're interested, but it's typical standard criticism versus a very strong epidemiological finding.

Frank Maddux: I think in our real-world evidence, we've seen patients both with great vascular access and those with catheters, all who can achieve certain convection volumes getting benefit from this. And so even in the CONVINCE trial, I think there were probably 15 or so percent of the patients with central venous catheters, and they too got the benefit.

Giovanni Strippoli: And there's this point that I always make. To do a trial like this is a monumental effort. It's extremely difficult. And the one thing that really is sad is to see that we don't rapidly action the findings of these trials when they become available, because we only focus on certain nuances, most of which are often wrong. So, one of the criticisms I hear, and I see also in the literature is, well, 80% of the patients in Convince had a fistula. So, the data are not applicable. And my answer to that as a clinician, not as a scientist is, well, this means we should do more fistulas. It doesn't mean that HDF doesn't work. It means we should do more fistulas. 

Another thing I hear very often is, well, there was a difference in the effect in people with and without diabetes. That is actually not true. It's absolutely not true. There is no interaction. The effect is exactly the same in the Convince trial. Of course, there is a figure in the New England Journal paper where the risk estimates look slightly different, but the confidence intervals overlap. And also, when we look at the totality of evidence and we did the meta-regression analysis where diabetes was the covariate, there's no difference. 

People always like to look and challenge the data, which is of course the nature of scientific debate. But I do think that there is a time when we have to action things because otherwise, we run the risk of what I call knowledge translation failure. This is very, very dangerous. I recently brought the example that when thrombolysis, was introduced for the treatment of myocardial infarction, there were trials done against aspirin. And once the trials were conclusive that thrombolysis was better than aspirin, people continued to use aspirin for another 20 years. And it took another 10,000 patients in the trials, which means half of them received the wrong intervention, until thrombolysis became standard care. 

So, I think we need to avoid this. And in nephrology, we have a long history of this. I mentioned erythropoietins, I mentioned calcium emetics, there is a lot of examples of this. And I think in this case, with dialysis patients, we really need to take this in our hands because these people are sick, they're not well, they deserve better. And if there is an intervention that is actually effective, both on survival and patient reported outcomes, and the totality of evidence is strong, we need to really action it. That's my personal view.

Frank Maddux: We've got an entire country that's a very large healthcare system in the world, in the US, that I have great interest in accelerating the adoption and socialization of HDF and ultimately moving it towards a standard of care. What do you think the work is that needs to be done practically either through policy, through socialization with our peers, other things that could be done right now to get to that more rapid actionable response to these trials.

Giovanni Strippoli: One of the things I have experienced in my life is that there is always this time gap. It takes time for people to really action these things. And the other thing that I have always been a little bit struck by is that there has been a lot of medical problems that have become engineered by governments, payers. So, the overall management of people on dialysis has often become more than a medical matter, an engineering matter, the pursual of higher, let's say adequacy because this is one of the parameters in the rating systems the pursual of X percentage of people who are receiving this type of treatment because that's what we can get reimbursed for. Now from a strictly medical standpoint this is really not the right thing because what we know is that patients will benefit from certain intervention and it's not a matter of subtle nuances. Like we said, the volume concept or the adequacy concept. HDF significantly reduces cardiovascular death. So, one of the things that I believe we need to do in the United States and in places where there is no uptake yet of the technique is to make the entire community in a way, I don't want to be offensive here, but I would say literate in terms of the science. This is very solid science. We're not talking here about one study that showed something that is potentially interesting. We're talking about solid, solid evidence. So, we have to make both the medical communities and the authorities, the payers, et cetera, totally literate about this. That's one thing I would certainly do and that is of course done with a lot of communication around the topic. 

The other thing is of course, and I've had some examples of this when I was working with Saudi Arabia, etc. It's important to make sure that within the actual agreements, in the United States it works in a certain way, but there is geographical variability like where there are tenders or where there are specific governmental agreements about the service to be delivered. This needs to be on the top. I gave you the example of Saudi Arabia just because I was heavily involved in some process of integration of clinics in Saudi Arabia. One thing I remember is that there were a lot of policies and a lot of deals and agreements around achieving X percent of people receiving a KT/V or more than 1.4 or a PTH within a hundred and five hundred picograms. 

This falls on the bottom because now what you should have really top, top, top line is where there is solid evidence. So, we would need to renegotiate these types of agreements where our sort of hard milestones would have to be the proportion of people who receive an effective intervention rather than the proportion of people who achieve a certain level of a certain surrogate biomarker, whether it's validated or not. So, these are two things. Of course, there would be many more. It's a difficult question and I'm happy to debate this more with you over time but it's this is going to be a challenge but certainly a worthy challenge for these highly deserving people.

Frank Maddux: We're here today with Dr. Giovanni Strippoli, who has been talking with us about this combination of using diffusion and convection as physical techniques in hemodiafiltration to improve the lives of our patients. Giovanni, thank you so much for being with me here. Any final comments?

Giovanni Strippoli: Well thank you Frank, it was a real pleasure for me to be here and as I said I do hope that also this activity has served the purpose of informing people about the totality of evidence around convection versus diffusion and hopefully this will generate a thorough debate and an update of a technique that certainly provides benefit.

Frank Maddux: Thank you so much and look forward to working with you and hearing your perspectives again as we move this conversation along, especially here in the U.S.

Giovanni Strippoli: Thank you, Frank.