Slide 1

Mr Chairman, Ladies and Gentlemen thank you for the introduction and also for the invitation. So this is my title, it’s a whole mouthful but actually the question is, is there enough evidence for the importance of removal of middle molecules? Is there a rational in trying to get a kind of target level where we should go underneath?

Slide 2

Now, this is the summary of my presentation. We first go to the old markers of dialysis adequacy, the good old Kt/V, which is as of today still the only adequacy marker which we are using. We’ll try to answer the question whether this Kt/V is really representative of what is happening with other molecules that induce uraemic toxicity. We will come to the core of this presentation which are the middle molecules and give some evidence or try to give some evidence about their toxicity and finally, we will talk about removal: are there arguments to say that by trying to remove organic molecules by high flux membranes there is a clinical benefit for our patients.

Slide 3

Now, Kt/V has been around already for many years, as you all know and there are quite a number of observational studies that show a survival advantage when Kt/V is becoming higher but as I’ve already said, most of these studies are observational and they go back to the end of the previous century which means that actually the level, the target levels of Kt/V are obviously lower than those supplied as of today and this is also illustrated in this slide from a study by Held and co-workers where you can see that the maximum target level corresponds to what we try to aim at, so the average, as of today. So there is some problem with these studies and their interpretation. We can’t turn back the clock, so it is impossible to do now control studies going back to the low Kt/V levels that were applied 10 years ago and for that reason probably they’re never will be a high evidence, a level proving this Kt/V is superior.

Slide 4

The same can be said of this study which goes back even to 1993 by Owen and co-workers but the interesting thing of this study, is that it is looking at 2 factors at a time. On the one hand, urea reduction ratio which roughly corresponds to Kt/V but on the other hand, serum albumin. From this study first it is clear that Kt/V has a positive impact on the outcome of these patients if it is rising but it’s going not very straightforward and the result is not very spectacular if we compare it to what is happening to albumin where the differences are much more straightforward and also more important. So, what you see with albumin is for each stratum a huge impact and a really linear impact which is very much less predictable and controllable for Kt/V. In other words, albumin which has been related to inflammation, to vascular disease and to malnutrition is telling something more and something more happening than what you control with Kt/V of urea. Or in other words, Kt/V urea is telling you something about the outcome of your patients but lots of other things are happening and these things are not necessarily under control with Kt/V and they have possibly something to do with a huge image that is related to inflammation, vascular disease and malnutrition. Actually, the same message comes out also from the only control study that has been undertaken in this area and for Kt/V urea I’m sure that this will remain the only control study for a long time or never ever someone will make the effort again to look at Kt/V in a controlled way.

Slide 5

You are very familiar with this slide, the conclusion is obvious for either standard dose or a high dose of Kt/V, there is no impact on survival of patients all over the place. We must point out that the standard dose here was somewhere around 1.4 single pool 1.3, 1.4 single pool, so it is high as compared to the previous study by Held. Whereas, the high dose was somewhere around 1.7-1.8, but again the HEMO study tells the same thing, the same story actually as the study from Owen and co-workers, namely that there is more than a urea small water soluble compound removal because in the other arm of the study there is a comparison between low flux and high flux membranes. In other words, membranes that have small pore size for the low flux or large pore size for the high flux and the latter are removing larger so-called middle molecules and although in this study at primary analysis there was no significant difference but there is a trend, whereas at secondary analysis, and I’ll come back to this later during this talk,there is significance for cardiac death and for the summation of cardiac morbidity and mortality. In other words, here also the story is told that there is more than removal of small water soluble compounds and that there’s something more that might have to do with removal of middle molecules.

Slide 6

Why is this? Well, if we look at the physical chemical characteristics of the compounds that are provoking uraemic disease, we can classify them along their characteristics which influence their removal by dialysis and there is a huge group of small water soluble compounds their molecular weight arbitrarily is set below 500 Daltons and typical prototypes are urea and creatinine. They are easy to remove with any type of dialysis whatever it is. It shouldn’t be high flux, it shouldn’t be convection. Also with standard dialysis there is already a decent removal. The other characteristic of these molecules is that not necessarily they are very biologically active, in other words they may be not very toxic and this in contrast to the two other groups which we can also classify as called difficult to remove molecules which can go out only with very efficient dialytic strategies and have a totally different kinetic behaviour as compared to urea and creatinine. In other words, Kt/V urea does not represent their removal pattern. These are the protein bound molecules, prototypes, phenols and indoles and the larger middle molecules arbitrarily above 500 Daltons, for instance, beta 2 M and cytokines and 500 Daltons is really a minimum since most of these molecules are larger than 10,000 D and so they are really big and you’re not removing them, if you do not open up the pore size of the membranes in other words if you do not go for high flux dialysis.

Slide 7

Now, what’s the toxicity of these molecules? Are there arguments that these middle molecules are really bad? There is, I’m going to focus only on one problem, it is the cardiovascular problem, which I think is the main thing that bothers us for the time being in nephrology and also in dialysis and together with the European Uraemic Toxic Work Group or EUTox 2001 already we produced this review article where we were looking at potential candidates for influencing the cardiovascular system in a negative way and we subdivided, we made tables of potential responsible molecules which were influencing the 4 major players in causing cardiovascular damage which are white blood cells, polymorphonucelars or monocytes-macrophages, endothelial cells, platelets and smooth muscle cells cells. So as you can see again, the potential culprits are to a large deal these so-called difficult to remove molecules, all the middle molecules are white here and the protein bound and water soluble are in yellow. Actually, there is only one water soluble compound it is oxalic acid.

Slide 8

So if we focus on the main uraemic toxicity problem but the same can be said of the other problems there are certainly arguments that the middle molecules based on bench studies, on in vitro studies are quite important culprits in provoking this disease. Time went on and from 2001 to later on we learned more about other molecules and this is let’s say a list of currently known middle molecules as of today and those in red have the potential to influence the vascular system. In other words this list is quite impressive and at least on the basis of these bench arguments there is reason to accept that we have indeed a benefit or we could have indeed a benefit by removing more of these middle molecules.

Slide 9

Now, the next question is then, is there clinical evidence to argue in favour of this middle molecule removal. Here we must say that most of the studies up to today have been focusing on high flux versus low flux, high flux being used in a dialysis mode not in a convective mode. In other words the optimum of removal has not yet been reached in these studies although there is some internal haemodiafiltration of course in all these open membrane structures. But in any way up until 2005 we had at least 8 or 9 studies that were evaluating the impact on survival of high flux versus low flux. Unfortunately, most of these studies are again epidemiologic, are observational and there are only 2 prospective studies around now. The remarkable thing is that all the observational studies give a difference but that the 2 controlled studies do not. Now, this Italian controlled study by Locatelli and co-workers is interesting in a way that in the conclusions the authors say that they find the study underpowered mainly because of the good clinical condition of the patients who were enrolled in the study. They suggested in the future if similar studies would be done, patients should be selected in function of comorbidities and have more heavily affected patients enrolled.

Slide 10

The other one is the HEMO study which we discussed already before and which we come back to. The HEMO study again at primary analysis did not show an advantage for the high flux but later on a subanalysis there was an advantage for cardiac causes of death. You can see below on this slide.

Slide 11

This was further extended in this other study by Cheung and co-workers in JASN. Here again it comes out that if they look at all patients together although in global mortality of which is here the topic of this slide there is no significant difference in mortality. If they look at cardiac death at first cardiac hospitalisation or cardiac death in the Magenta squares, you can see that there is a survival, a global survival advantage. The other thing is that if they select patients who have been more than 3.7 years on dialysis, at the start of the study they also have both a global survival advantage and a survival advantage on the cardiovascular front. Now, about this 3.7 I have to give a small explanation the HEMO study indeed was a study which was done on prevalent patients, that means that the patients were already on dialysis when the study was started which is also a potential disadvantage of the study because there is a risk for carry-over effect and influence of previous dialysis strategies that the patients underwent at the moment of enrolment into the study. But apart from that as a consequence the median time on dialysis was 3. 7 years at the moment that they were enrolled in the HEMO study and when they picked out those who had been on dialysis for more than 3.7 years, so who had already been a long time on dialysis then obviously in this subgroup there was a survival advantage for the high flux membrane which might be that somehow high flux is taking away molecules which under normal conditions at the beginning of haemodialysis are taken away by the residual renal function but as this residual function resides that high flux membrane may take over what is not taken over by low flux membrane.

Slide 12

Other subanalysis of the HEMO study again showed a similar thing. This is a recent one by Delmez and co-workers focusing on cerebral vascular disease and again high flux is superior to the low flux membrane. Another study which was very recently published is the one by Kramer and Wanner in the American Journal of Kidney Diseases.

Slide 13

This is a secondary or subanalysis of the 4D study in which as you know has been undertaken to evaluate the effect of a statin on survival of German dialysis patients with diabetes and as you know, this statin turned out to be non-effective, at least at global analysis but here there was enough data available to compare also different dialysis membranes and as you can see, the high flux membrane here in red is largely superior to 3 different strata of low flux membranes where among these types of low flux membranes there is still some difference which in my opinion can be attributed to the biocompatibility of these membranes because low flux cellulosic is at the lowest level of survival. In other words, this study also tells the same story but again this is a secondary analysis study which has been developed to to search for something else. This study tells again that at least in diabetics there is a superiority of the high flux membrane.

Slide 14

That’s all about comparison between high flux and low flux. There is one other interesting recent analysis based on the HEMO study but which contains no selection of patients according to the membrane, so these are all the patients taken together and then Cheung et al. looked at the relationship between beta-2 microglobulin levels and outcome where beta-2 microglobulin in this case means a molecule representative for the other middle molecules but where the concentration can be at a certain level whatever the type of membrane used. Look what happens in this study. When we go from lower to higher beta-2 microglobulin levels, there is obviously a negative impact on survival or the mortality risk is increasing progressively. Their beta-2 M levels are relatively high which in my opinion is due to the fact that the high flux which is included in this study is not a real high flux but is actually a medium flux which is due to the American standards which are different to the European standards which in turn have to do with the quality of their dialysate. But anyway it means that this shows that the more removal we have, the better things go but if with this reflection that this is their lowest point, 27.5, so nobody knows what happens below here, is this curve still getting lower? What is the answer to the question what to do with middle molecules? Is the answer the more you remove the better or is the answer, is there a certain threshold below which we shouldn’t go? That is not answered in this and in my opinion this is the question which we should answer in the future.

Slide 15

Is there something about convection? Well, there is very little and what there is, is again observational. It’s this large analysis of the DOPPS database by Bernard Canaud also recently in Kidney International where it is observed that as compared to a low efficiency haemodiafiltration which means that at an exchanged volume, a volume ultrafiltered and returned to the patient as a pure saline substitute, 5-14 L that there is a gain in survival if shifting between low to high efficiency where the exchange volume is 15-25 L and this gives a survival advantage of 25%. However, this huge amount of volume can only be obtained in the setting of online haemodiafiltration which means that there is one caveat, the quality of this dialysate returning into the blood stream of the patient should be optimal. We have some data that this is not always the case it will be presented as a poster at this meeting on Saturday I think by Doctor Glorieux from our group. Indeed, we overestimate the quality of our dialysate which is due to the fact that the tests we used, the bacteriological ones, the LAL tests are missing a number of contaminants and we should be aware of this and I think we should aim at the optimum quality of the dialysis water to do this online hemodiafiltration.
Are there any controlled studies? There are two running, one in France and one in the Netherlands. As far as I know for the Dutch study, at the moment it’s still going on so we will know the results only in a few years and before that I think that there is a suggestion that more convection is good but there is not more than a suggestion and we are anxiously waiting for the results.

Slide 16

Is this is all that can be said about this topic? No, there is one information missing. It’s the MPO, study the membrane permeability outcome study which was a large European multicentre study. Again, in dialysis mode high flux versus low flux, what can we say about the MPO?

Slide 17

Nothing because the results are under embargo, they’re still under embargo till tomorrow afternoon, so instead of showing the results, I’ll show you the steering committee but nevertheless, I want to point to a number of important differences between the MPO and the HEMO study.

Slide 18

The HEMO study then being our main predecessor and many of these characteristics I think are in dysfavour of the HEMO study, so that we can expect that we could learn something more. Something is not mentioned here that is that the MPO study is essentially focusing on hypoalbuminaemic dialysis patients with a serum albumin below 4 g/dL which is important because this answers the question of the comorbidities, which I pointed out at the beginning when we were discussing the Locatelli study. There is in the MPO study only 1 arm. There are two groups, high flux versus one low flux which makes the design simpler versus the HEMO where there were 2 arms and 4 groups and makes it easier to reach power. MPO more than 660 patients actually somewhere around 740 compared to 940 in the HEMO study. So, for two arms this number is obviously relatively lower than the 700 for the MPO study. Here’s the follow up time for the MPO from 3-7.5 years compared to 1-5 for the HEMO study. That type of patients is incident in the MPO and prevalent in the HEMO and I’ve already criticised the enrolment of prevalent patients because of the potential carry over effect of previous dialysis strategies. Treatment time according to European standards is set as a minimum one can go as high as possible compared to the HEMO study where the maximum is set and one can go as low as possible and conform with the American practice where dialysis time is kept as short as possible which is obviously not advantageous and we think that dialysis time within Kt/V has an effect on the clearance. There was no reuse in the European study and a definition of high flux was much more liberal which means that in the HEMO study the high flux was actually perhaps something more than middle flux.

Slide 19

So in other words I think that it is very interesting to look at the data that will come out in the HEMO study that will be presented tomorrow afternoon in the session of ongoing clinical trials by Doctor Locatelli and this will happen between 15:25 and 17:50, so you are all cordially invited to attend that meeting.

Slide 20

In summary, we discussed the role of Kt/V which might be a good baseline standard but which is not representative of many things, there’s more happening in uraemic patients and in dialysed patients. There’s something else and this something else is probably or potentially related to the middle molecules, there are a lot of bench arguments that point to their toxicity and the next question is of course, about clinical arguments of their removal. There are a number of studies about high flux and definitive evidence is lacking as of today, perhaps no more tomorrow. You can see the additional role of convection is very suggestive but has to be proven as well. But indeed I think that this data altogether point to the fact that we should look for markers in the middle molecule area and that this will bring us further in our effort to improve the quality of life for dialysis patients. I thank you for your attention.