CME Slides Forum - C.A. Hutchison

A joint Congress by ERA-EDTA and ISN

NEW MEMBRANES

Colin A Hutchison, Birmingham, UK

Chair: Claudio Pozzi, Milan, Italy

Pierre Ronco, Paris, France

Dr Colin A. Hutchison
Department of Nephrology
University Hospital Birmingham
Birmingham, United Kingdom

As we’ve heard already renal impairment is a frequent complication of patients with multiple myeloma affecting up to 40% of patients at presentation of disease depending on the definition of renal failure we use. Unfortunately of all myeloma patients about 10% will require dialysis support at some stage. This renal failure is associated with greatly increased morbidity and mortality of this population. However, several studies have shown that if renal failure is reversed the patients’ outcomes dramatically improve really making the treatment of renal failure in patients with myeloma a priority in their management.

Now when we focus on those patients with myeloma the and severe renal failure the renal pathology is almost always cast nephropathy. The co-precipitation of the monoclonal FLCs in the distal tubules with Tam-Horsfall protein. So a direct consequence of the very high concentrations of monoclonal light chains in the serum, surely making targeting, reducing those levels an appropriate target in the management of these patients.

This concept is supported by this data from Nelson Leung and colleagues at the Mayo clinic last year which shows that patients with renal recovery on this graph as renal responders with cast nephropathy are those patients who generally achieve a greater than 50% reduction in these serum light chain concentrations early in their disease management, highlighting that this is an appropriate target.

Of course, that’s the logic behind plasma exchange which we’ve just heard so much about.

So why is it that this logical treatment has failed to result in positive control trial results? Well, that’s the starting point for our work in Birmingham, to really understand why plasma exchange doesn’t work in this context we designed a two-compartment model with the help of the department of engineering from Warwick university.

Compartment 1 represents the intravascular space. Compartment 2 the extravascular space. Light chains, of course, are produced into the blood but because of their small molecular weight they redistribute fairly easily between the 2 compartments. The result is therefore, that at any one time approximately 80% of light chains are extravascular simply because of volume distribution.

Now in health light chains are removed very rapidly from the kidney giving them serum half life of between 2-3 hours. But in the context of renal failure this clearance becomes minimal and removal by the reticuloendothelial system becomes a predominant mode of removal and this is a much slower process increasing the half-lives of light chains up to 4-6 days dramatically changing their kinetics.

Now what this model allows us to do is to simulate how different modalities of removing light chains from the serum, such as plasma exchange, would affect both the intravascular concentrations but also the total body load of light chains.
Now, we saw one of these graphs earlier but just to run through the basics with you this is a hypothetical patient presenting with 10.000 mg/l of free Kappa. That’s a lot because in comparison normal ranges are up to about 25mg/l but in the context of myeloma and severe renal failure that’s not an uncommon value. We then have the treatment period in days along the x axis and the curve here is how the patient’s serum concentrations would fall away if we managed to stop production on day 0. So the perfect chemotherapy killed all the tumour on day 0 and then removal was by reticuloendothelial system alone. Clearly an unrealistic scenario because no chemotherapy will ever achieve that.

So what would happen if we had effective chemotherapy and were able to kill 10% of the tumour per day? Well, you can see that the area under the curve is dramatically increased and it’s the area under the curve that is so important because this represents the ongoing load of light chains on the kidneys as they try to recover. So we can understand for the first time why patients with cast nephropathy and on dialysis don’t recover their renal function it’s simply because they’re exposed to very high levels for a long period of time. From animal models we know nephrotoxic levels are about 500 mg/l and you can see on this graph that the patient is exposed to that concentration for nearly one month. Again, rat models have demonstrated that you have a window of approximately one month to reverse a cast nephropathy before you get irreversible damage setting in.

So what will plasma exchange do? Well plasma exchange removes very effectively whatever is in the blood but because of its short duration it doesn’t allow the extravascular compartment to be cleared at all. So despite a rapid drop when the patient receives plasma exchange, the actual area under the curve is not changed dramatically. Clearly if you gave the patient more intensive perhaps daily plasma exchange, you could reduce this area under the curve but not dramatically.
So with our understanding from those simulations that the main failing of plasma exchange was its short duration not allowing the extravascular compartment to be cleared, we evaluated hemodialysis because clearly we use this regularly on extended therapies. Now the limitation of hemodialysis was always going to be whether the current dialysers could actually remove light chains. Just to remind you our current high flux dialysers have molecular weight cut-offs of generally between 10-15 KD. Their light chains are classified as middle molecules with molecular weights of 22.5 and 45 KD respectively, so well above what these conventional high flux dialysers should be able to remove. There is however, a new class of membranes called protein permeable membranes or superflux dialysers that have higher molecular weight cut-offs potentially in the region that would allow light chains to be removed.

So using an in vitro study, using an ultrafiltration model we looked at whether these different dialysers could remove light chains. The initial results were encouraging. When you looked at the reduction of serum light chain concentration, it would appear that all dialysers could remove light chains, a slightly surprising result.

However, then when we looked at the proportion of light chains that were coming out into the ultrafiltrate we saw a great variability, particularly when we focused on the superflux membranes.

Just to emphasise that both the Torays which are PMMA based membranes and the gamma rates are 1100 all resulted in significant reductions in the serum concentrations but look at the dramatic differences in what comes out in the ultrafiltrate, really emphasising that a lot of these membranes are very effective in absorbing proteins on the surface of the dialyser such as the Toray membrane but only one was allowing light chains to effectively come out into the ultrafiltrate. This is a very important observation because it represents the capacity of the system to remove light chains. If we’re removing light chains by simply binding them to the surface of the dialyser, we’re limited by how much protein can bind to that surface. Generally that’s between 1 and 2 g. Remember these patients tend to have tens or hundreds of grams circulating in their serum at any one time making absorption not a suitable modality for these patients.

So what is it that makes the high cut-off dialyser unique?
It’s simply the sizes of the holes, they’re bigger. What we have here is a pore distribution for your high flux, high cut-off and a plasma filter. By this slight shift in the pore size, the high cut-off dialyser can remove a whole different group of molecules classified as middle molecules. The pictures on the left show the entrances to the pores and just how visibly different these are.

So we then undertook a series of in vivo studies to try to optimise light chain removal by hemodialysis. Some of our key observations were firstly, despite no differences in molecular weight the clearance rates achieved for kappa and lambda were sufficiently comparable to mean that you don’t have to treat the patients differently. As expected, we found that clearance of light chains being a middle molecule was increased when ultrafiltration was used. Most importantly we found that despite some loss in clearance rates over the duration of an extended session the procedure remained effective for up to 8 hours. One of the limitations we found using this dialyser because it wasn’t designed for routine hemodialysis it has a small surface area of 1.1 m² we therefore started to use routinely 2 dialysers in series which not only increases the surface area but also results in a convective circuit where light chains and water are driven out of the first dialyser and there’s a re-dilution effect in the second dialyser greatly increasing the clearance rates. One of the most important things that we’ve observed is that because of the serum coefficients of this dialyser there is significant albumin loss which must be replaced at the end of every dialysis session, without this the patients would become markedly hypoalbuminic.

So going to back to our simulations. This red line now represents what would happen to a hypothetical patient if we removed the light chains by dialysis for 4 hours 3 times a week. Already significantly better than plasma exchange. That’s simply related to the fact that a dialysis session is approximately twice the length of a plasma exchange session. But of course, we wanted to be able to give an extended therapy and this black line represents what would happen if we had 12 hours of hemodialysis on a daily basis. Potentially for the first time dramatically reducing serum concentrations of light chains and possibly giving the kidneys time to recover.

So with this knowledge we were then able to get an ethical approval to run a pilot study of this new treatment in patients with multiple myeloma and biopsy proven cast nephropathy. All the patients we studied were dialysis-dependent. The principal outcome that we wanted to show was that it is possible to actually achieve a reduction in serum light chain concentrations. We secondly described the outcomes for this population and then compared them with a case matched control from Birmingham.

During a 2-year period we screened 27 patients, 19 of whom met our inclusion criteria. All the patients were treated with combined protocol which included chemotherapy and light chain removal by dialysis. The chemotherapy we used was thalidomide and dexamethasone for patients with de novo disease and bortezomib based therapy for those with relapsing disease.

The dialysis schedule we used was 8 hours daily for the first 5 days and then reducing to alternate days through to 21 days or stopping sooner if there was renal recovery or the light chains dropped below the proposed nephrotoxic range of 500 mg/l.

These are the results of the first patient we treated. He’s an Afro-Caribbean gentleman in his early 60s with a known IgG kappa MGUS who presented with cast nephropathy. What I’d like to point out first is he had a very low level of his kappa light chains at just over 1g/l emphasising this individual light chain had a higher affinity for Tamm-Horsfall protein and precipitated out at a low level. Secondly, I’d like just to point out that every dialysis session is two dots joined by a line. So every session resulted in a reduction of light chain concentration but as we saw earlier with plasma exchange there’s frequently a rebound in between sessions. This rebound is due to this redistribution of light chains back in from the extravascular compartment and secondly ongoing production. Now in retrospect this patient had a very sensitive tumour to dexamethasone as after just his second pulse of dexamethasone he essentially stopped producing light chains and that’s why we see the levels rapidly dropping away. Also by this stage his native urine output had increased dramatically and he was therefore, clearing light chains on his own. This is over 2 years ago and he’s remained dialysis-independent since about day 21. He went on to have a bone marrow transplant with a GFR of about 45 ml/min making his management much easier for our haematology colleagues.

Now when we look at the primary outcome for this population of light chain reduction, we found that the group fell into 2 groups. Firstly 6 patients hit early complications principally infections that required their chemotherapy to be withheld for a variable period of time. 13 patients who went through their treatment course uneventfully. Very importantly we found that patients who had their chemotherapy stopped or withheld never achieved a reduction in their light chain concentrations. Those who had continuous chemotherapy and dialysis all achieved a sustained reduction emphasising that no matter how much light chains you remove unless you -- with effective chemotherapy the light chain concentrations in the serum will remain raised.

When we looked at the secondary outcome of renal recovery in this population, 14 of 19 patients became independent of dialysis, far above that expected from the published literature. That was the 13 patients who had a sustained reduction of the light chains and then one patient from the 6 who had a later reduction in the light chains.

When we compared this to a case matched control population from Birmingham, the difference was very significant. It’s important to represent that not only did these patients have the benefit of dialysis but of course, chemotherapy has changed in the years between these two populations.

As expected, survival related to renal recovery, patients who became independent of dialysis had a far better survival than those who remained dialysis-dependent.

Just to recognise the limitations. This is an uncontrolled study with small patient numbers and we’ve only assessed this procedure in patients with cast nephropathy and dialysis-dependent renal failure.

For that reason we’re running an RCT now in Europe, in Germany and in the UK to further evaluate this and in the GAMBRO symposium at lunchtime there’ll be more details of this presented.

So in summary, we know that high cut-off hemodialysis can remove very large quantities of light chains. But without the combination or with the effect of chemotherapy that won’t result in a sustained reduction in light chains. We do however, know that achieving an early reduction in light chain concentrations will result in improved patient outcomes in terms of renal failure or renal recovery and overall survival.

Many centres worldwide have now taken up this treatment. Initial results are encouraging. As I mentioned, we’re running this RCT which we should have recruited and completed early next year.

Finally, just to thank my colleagues in Birmingham who have helped this work possible.

Chairman: Thank you very much Doctor Hutchinson for this cutting-edge presentation and for pointing out that not only the amount but also the quality of the light chains and their propensity to interact with Tamm-Horsfall protein are very important factors of progression or regression of renal disease. The paper is now open for discussion.

Question: If you quantify the amount of the light chains removed compared with this method and plasmapheresis is there a significant difference between the two?

Question: The question is how long is the renal recovery sustained after successful treatment?

Dr. Hutchison: One of the interesting things we’ve seen is actually these patients carry on improving renal function for about 6 months afterwards and largely that’s related to ongoing response to chemotherapy, drop in their light chain levels further but also just a natural recovery in the kidney.

Question: Thank you, a really clear presentation. Two questions. If it’s just a question of duration why not just do plasmapheresis for longer and not bother with the new technology? But if you’ve got a good answer to that one, you mentioned very briefly the need for albumin replacement but a membrane cut-off of that size must also remove many other molecules. Are there any other possible unintended consequences of high cut-off dialysis?

Dr. Hutchison: To address the first question the benefits of dialysis over plasma exchange is firstly the safety in that you can give a lot more dialysis than plasma exchange and more regularly because although there is a higher cut-off removing molecules up to the level of albumin, you’re not depleting the patient of intact immunoglobulins for example. Also there’s a straight practicality advantage that this is much easier to perform in the dialysis unit, it’s essentially standard dialysis to the nurses and they become rapidly very comfortable with it. Sorry I’ve forgotten your second question.

Dr. Hutchison: Yes well we actually didn’t have to do many of those studies because this dialyser was initially designed for removal of IL-6 in sepsis patients and those early studies demonstrated safety in terms of clotting factors and we’ve not hit any complications with our patients.

Chairman:Other questions? Doctor Hutchinson, what population of patients would you target for dialysis with the membrane? Only those who require dialysis or also patients with severe renal failure but not requiring dialysis?

Dr. Hutchison: I think it’s a very interesting question. The removal of light chains is very closely linked to clearance by the kidneys and in a study of polyclonal light chains we found a dramatic difference once you get a GFR of less than 30 ml/min that their serum half-lives dramatically change. So far we’ve just studied dialysis-dependent patients because of the intensity of the regime and possible side-effects but in the longer term undoubtedly some clinicians will dialyse patients with GFRs of between 15-30 ml/min and it may well prove to be beneficial.

Chairman:Thank you other questions? May I add a clinical comment? Those strategies are very important and very promising but it is very important first to apply symptomatic measures that is rehydration, correcting hypercalcemia, avoiding the use of non-steroidal anti-inflammatory drugs and others etc. etc. which should prevent the aggravation of renal failure or even the appearance of renal failure in patients with myeloma cast nephropathy. I think it’s very important to remember that before the use of plasmapheresis or the use of dialysis membranes such as HCO, symptomatic measures should be applied to patients what’s your opinion?

Chairman:Thank you I would like to thank all the speakers for really brilliant presentations and the audience for