Introduction
Donal O'Donoghue, MB, ChB: Hello. I'm Donal O'Donoghue. I'm a renal physician in Manchester, practicing at Safford Royal Foundation Trust. I'm also the National Clinical Director for Kidney Care in England. It's a pleasure to be with you for this Skype chat. My colleague, Lynda Szczech, is going to join me. We are going to talk about some of the challenges we face in the clinic in terms of when we should initiate dialysis. There have been a lot of publications about this over the years, and I'm not sure we're that much further forward. There is a great study published in the New England Journal of Medicine from our Australian colleagues, the IDEAL Study, which I think includes about [800] patients randomly assigned to an early start or a late start, and [it separates] the groups, but not quite as much as they may have intended. The analysis didn't support early startup. Did you see that paper Lynda?
Lynda A. Szczech, MD, MSCE: Oh, I sure did, and wasn't that the talk of all of the meetings last year. So, hello, everyone. My name is Lynda Szczech. I'm a nephrologist at Duke and president of the National Kidney Foundation this year.
The IDEAL Study and Early vs Late Dialysis
Dr. Szczech: [The IDEAL] study really answered so many questions, because, as you know, when to start dialysis traditionally was a qualitative discussion with your patients. How do you feel? Are you emotionally ready? Are you physically having symptoms? [With] that sort of qualitative discussion you've got to wonder if you're doing it right. But when they randomly assigned these 800 odd folks, and they found there really was no difference in mortality and maybe some of the point estimates for heart events favored the later start, it made you feel better about waiting for symptoms to occur and for the patient's body to tell you, rather than the numbers, that it was time to start. Is that what you took away from it, Donal?
Dr. O'Donoghue: Yes, I did. I know the population was somewhat younger than our population in England, but it was representative of Australia and New Zealand, so I think that's what we want to see. I took away from it that early start doesn't necessarily confer. I think that the later-start patients were started for uremic symptoms, whatever they may be. I guess that's what we've been doing traditionally. In one sense it reinforces the way we've been doing things.
Perhaps it asks a larger question: what other things do we need to measure to know when starting dialysis or preemptive transplantation, which is obviously much rarer. I think that gets into a very different place than the traditional biological approach with the relatively easy measurements, ie, diabetes, age, blood pressure, and comorbidities, but it's a fantastic study. Some people would say it's negative. I think it actually gives a very good platform. What did you think about the relevance of that population to a US population?
Dr. Szczech: They have that really nice forest plot where they look at the subgroups, and there was really very little difference between them. Even though [the study population] was a little younger, which is what you would expect with a volunteer bias that goes into clinical trials such as this, I thought that it probably was fairly generalizable.
Dr. O'Donoghue: I know there was criticism about the fact that, in the end, the separation between the 2 groups was I think 2.2 mL/min, per 1.7 m glomerular filtration rate (GFR). But there was about a 6 months' difference in start time, which I think many of our patients would appreciate. Again, it's conjecture on my part, rather than something I've got evidence on. I don't think many people would be standing up saying," I'd really like to start dialysis earlier than I need to." It's back to that question, when does it confer benefit? We know from clinical practice that starting too late is problematic, but it appears starting early in a planned fashion isn't the right answer either.
Dr. Szczech: I almost expected that there would be a higher rate of infectious complications, given the fact that there would be vascular access earlier in the early starters -- potentially some more catheters. But the twist that there wasn't actually increased infection risk, but [there was] increased cardiovascular risk was very interesting to me. As I started to think about the risk and the benefit of more clearance, while intuitively you'd think that more clearance had to be better, with that benefit comes the risk of delivering that clearance, and that risk might be cardiovascular.
Dr. O'Donoghue: Yeah.
Dr. Szczech: This is consistent with observational work that you were mentioning earlier, by Seth Wright.
Dr. O'Donoghue: There's a great study by Seth Wright and colleagues, looking at United States Renal Data System (USRDS) data, following patients from 1995 to 2006, so [about 890,000] incident patients and controlled for all the variables that you normally control for. [There was] a whole range of subanalysis, and the message was the same in all of those 4 groups: people started with a GFR greater than 15, 10- 15, 5-10, and less than 5-10 mL/min. They used the 5-10 mL/min as the standard in which to compare the others. There was a gradation in all the analyses from the greater than 15 mL/min (worst survival outcomes) down to the less than 5 mL/min (always better outcomes). Those with less than a 5 mL/min GFR were younger, had less diabetes, and less comorbidity. But really, as you would expect with those numbers [there was] a very clear statistical difference. For all analyses there was a consistent message -- consistent statistically, and with clinically significant results -- showing that the earlier the start, the worse the outcome, right down to less than 5 mL/min.
So this is food for thought there. The groups were different and it was observational. One can't draw firm conclusions from that, but if you put these observational data from the United States together with the randomized control trial from our Australian and New Zealand colleagues, I think there's a very clear message. We don't have evidence that an earlier start, when presymptomatic, benefits them in terms of survival. We've actually got reasonably substantial evidence that harm is done, in comparison to waiting to when people need to start clinically.
So what do we think of that? What are we taking to our day-to-day practice from that? Well, the Australian study was of people who had enrolled in a study, and therefore, were prepared for replacement therapy, and that's not necessarily the norm in our various countries. We'd like it to be the norm -- that's the sole thing about preparation. There's something about what we aren't measuring here, to know when to start. I think you said at the beginning, Lynda, we're all a little bit uncomfortable about making decisions based on the absence of evidence. When we find ourselves in that situation it behooves us to stand back from the evidence we've got, and say are we asking the question the right way.
How Often Should We (Can We) Dialyze?
Dr. Szczech: It brings up the big picture; what do we really know about the delivery of dialysis. We know dose matters, and that started out with Ed Lowrie's early work in the 80s with the NCD trial and continued with multiple trials done by fabulous investigators. Now we're not quite sure when to start. At least early start doesn't confer benefits, and could potentially confer risk, so then how frequently [should we dialyze?] I'd love to get your take on the recent paper by Rob Foley in the New England Journal of Medicine, where he presented data that after the 3-day weekend, there was a greater risk for morbidity and mortality. What did you think when you saw that?
Dr. O'Donoghue: We've seen smaller studies before that hinted at the long break being problematic. However the power of the USRDS and the Rob Foley/Al Collins work is always high quality and large numbers are considered. I think it's a challenging paradigm of thrice-weekly dialysis. I know there are units that try and offer alternate-day dialysis. The logistics of alternate-day dialysis in a hospital-based environment are considerable. I'm looking forward to seeing what people say at the American Society of Nephrology (ASN). How are we going to react to this study? It takes in the daily dialysis regime (dialysis information that's arising from home dialysis), but also in-center. We need to share the information with our patients. We need to consider how we might take that observational data into some more randomized controlled trials. I don't think Rob Foley's paper is telling us we need to change immediately; it's telling us we need to think about what we're offering and how we're offering it.
Dr. Szczech: From a purely mathematical standpoint, it really is fascinating. I spoke with 2 or 3 reporters after that study came out through the National Kidney Foundation, and it was hard to explain to them the concept of dialysis dose, how much, and then how it was delivered and when. When you think about these 3 groups of papers (the dose trials, Ed Lowrie's NCD study, the hemo studies) and then the frequency trials, the frequent dialysis network, and now this trial about timing that thrice weekly might not be enough, we have an idea of the total dose but we don't really know how to deliver it, when to start, and then how to spread it out over the course of the week. Should this be done slowly in little chunks or in bigger chunks 3 times a week? It looks like the preponderance of the evidence is really moving toward slowly, more frequently, and in smaller chunks.
Dr. O'Donoghue: What's interesting is to go back and look at the very early reports. Why did we end up with thrice-weekly dialysis? I'm not an expert there, but I'd love to see somebody go back and look at the history. My understanding is that we arrived at thrice weekly because [if it was done] once a week, everybody died quickly. [If it was done] twice a week, I'm trying to think of a polite phrase for it, the patients didn't feel well at all. Again, there was high mortality. Thrice weekly was thought to be the bare minimum.
Certainly in the United Kingdom, our Renal Association guidelines say thrice weekly is the bare minimum. Fifteen years ago there were still people on twice-weekly dialysis, and because of resource restraints, it was something we had to battle as a community to put right. The evidence based on thrice weekly as the standard is now being challenged. The technology that is arriving allows us to practically be able to offer more frequent dialysis, or slower dialysis, or nocturnal dialysis, but the logistics of our system, because we've set it up in a thrice weekly basis, is clearly a significant challenge there.
Dr. Szczech: It most definitely is. The reporters actually asked me why we dialyzed 3 times a week. I didn't have a good answer for them. One suggested that it was because everybody wants to have a day off. It's unfortunate that our week has an odd number of days. I think that's the bottom line. If a week had an even number of days this would be far less of an issue. You spoke about resource constraint; that is really a delicate dance that I had this conversation with them about. If we can draw cause and effect from the Rob Foley article and we then take the frequency dialysis trial and extrapolate what seems to be a valid intermediate outcome of cardiovascular geometry to long-term outcomes, and certainly quality of life is a valid outcome, then we should be offering everyone a fourth treatment a week, at the bare minimum. The issue is really that of resource constraint. You need to demonstrate medical need, and how do you document better survival as a medical need?
Dr. O'Donoghue: I think it's logistics. The resources are really quite difficult, although I'm absolutely certain we should be sharing information with our patients. [We should be] exploring what modality of replacement therapy they would like, and within that the home environment. The home programs are small in both our countries. Smaller even in Australia, where it's probably the most common. There's no problem in offering 4 or 5 times a week dialysis. It may be that we should be offering that flexibility now that technology has given us smaller machines that are at least transportable, if not necessarily portable. That gives you freedom as well. Home hemodialysis has been kind of a niche area, and some would say that in the states even peritoneal dialysis has been a niche area.
Maybe we need to go back and revisit the offer and planning of dialysis straight away in the predialysis phase. As well as sponsor some trials or seek to get some support for some trials [for dialysis] around 4 times a week, or 2 in a day, vs thrice weekly in a hospital environment.
There's a kind of a problem, isn't there, of just one unit that says you net fabulous result. I call it the Tassan factor. Is it the dialysis, is it the people, or is it the fabulous cuisine of that part of France? So I make a joke. The point I'm trying to make: is it spreadable and is it sustainable? I think we need to have some high-quality randomized trials within the general environment rather than the niche areas, or the areas where people are particularly leading edge. Do we need to change the system around? And if we do, that's a big change.
Conclusion
Dr. Szczech: I think I have the perfect analogy to end our conversation with. We need to discuss with third-party payers why we don't provide 7 gm of vancomycin once a week to someone with normal kidney function, because 1 gm once a day or 1 gram twice a day kills the bugs and minimizes the toxicity. So maybe the analogy is the same with our "dose of dialysis." I put this in quotation marks. We should think about it in the pharmacokinetic sense, that we shouldn't try to deliver it all at once, or in big large chunks. But there is some benefit towards minimizing the risk and maximizing the benefit in slowing the rate of delivery.
Dr. O'Donoghue: That's a great way to end. When we're planning these trials I hope that we're also going to look at the quality of life in measures and that we look at issues of preparation. We [are also] going to look at a patient-centered approach to it, as well as the survival and vascular events. I think then we'll get a rounded picture of what the individual with endstage kidney disease can expect to see, and what we can truly offer for the changing times we're living in.
Dr. Szczech: Absolutely. Well, it's been a pleasure.
Dr. O'Donoghue: Bye now.