What form of renal replacement therapy should clinicians use for patients in the intensive care unit (ICU)? New research has connected the renal replacement therapy choice with mortality end points and renal replacement therapy dependency in patients with acute kidney injury. In this podcast episode, experts discuss their research in this area.
Ron Wald, MDCM, MHP, professor of medicine at the University of Toronto, discusses his article, “Initiation of Continuous Renal Replacement Therapy Versus Intermittent Hemodialysis in Critically Ill Patients With Severe Acute Kidney Injury: A Secondary Analysis of STARRT-AKI Trial,” published in the November 2023 issue of Intensive Care Medicine.
Jay Koyner, MD, professor of medicine and director of the nephrology intensive care unit at the University of Chicago, discusses his article, “Initial Renal Replacement Therapy (RRT) Modality Associates With 90-Day Postdischarge RRT Dependence in Critically Ill AKI Survivors,” published in the August 2024 issue of Journal of Critical Care.
This podcast is sponsored by Vantive U.S. Healthcare. Vantive supports true patient-focused treatments with industry-leading CRRT technology and is a partner dedicated to optimizing your clinical success in treating patients with acute kidney injury. Our commitment to you starts with education and provides complete support every step of the way. Visit us at vantive.com.
Transcript:
Announcer: This podcast is sponsored by Vantive U.S. Healthcare. Vantiv supports true patient-focused treatments with industry-leading CRRT technology and is a partner dedicated to optimizing your clinical success in treating patients with acute kidney injury. Our commitment to you starts with education and provides complete support every step of the way. Visit us at vantive.com.
Dr. Madden: Welcome to the Society of Critical Care Medicine podcast. I’m your host, Maureen Madden. Today I’ll be speaking with Dr. Jay Koyner and Dr. Ron Wald about AKI and clinical evidence for optimizing patient outcomes. New research has connected the renal replacement therapy choice made in the ICU with mortality end points and renal replacement therapy dependency in patients with acute kidney injury. Dr. Koyner is a professor of medicine and the medical director of acute dialysis at the University of Chicago. Dr. Wald is a professor of medicine at the University of Toronto and works at St. Michael’s Hospital. Hello and welcome. Do you have any disclosures?
Dr. Koyner: Yes, thanks for having me today. I do have some disclosures that are pertinent to this. I have received research funding in the past from dialysis companies, including both Baxter as well as Fresenius Medical.
Dr. Madden: And Dr. Wald, do you have any disclosures today?
Dr. Wald: First of all, thank you so much for inviting me. I have received research funding as well as speaker fees from Baxter. I’ve done consulting work for Alexion and sit on the scientific advisory board of Elias Therapeutics.
Dr. Madden: Very good. Thank you. Just to get to know each of you a little bit better, I want to start with you, Dr. Wald. Can you tell me a little bit about your background and interest in this type of research?
Dr. Wald: Thank you so much. When I became a nephrology fellow, I was always intrigued, particularly by the consults that we were doing in the intensive care unit for acute kidney injury, and it was evident at the time that we had very little evidence for a lot of the decisions that we were making. One of the questions that frequently came up was, what form of renal replacement therapy should we use for our patients in the ICU? The traditional approach that I was raised with, so to speak, was that we should use continuous renal replacement therapy and focus it for individuals who are hemodynamically stable. For others, we would use intermittent hemodialysis.
But as time went on, the excitement about CRRT was somewhat subdued due to clinical trials that showed no clear mortality benefit. But nonetheless, I had a strong feeling that CRRT was still an important modality for the right patients and thus felt that, are there any other important benefits to be gained by delivering CRRT and choosing it over other forms of renal replacement therapy? That led to a body of research and ultimately to some of the studies that we’ve published in this area.
Dr. Madden: And Dr. Koyner?
Dr. Koyner: I think for me, similar to Dr. Wald, some of it relates to just the clinical care that we provide and seeing in the past, there were patients who just did not tolerate or did not do as well on intermittent dialysis as they did on continuous renal replacement therapy. Over the years, certainly there are hard indications where people should require CRRT, but there’s a lot of uncertainty. As Ron points out, the data about some of the outcomes have always been mixed, to say the least, so that it becomes important to understand if we have this tool, and it is both labor and perhaps cost intensive, are there actually benefits to it, recognizing that there are lots of hospitals around the world where they don’t have CRRT? And is that putting patients at risk, or is that okay?
Dr. Madden: It’s a very good question to pose. I have to say, for myself, I’ve been doing this long enough that I’ve seen an enormous change in the technology. When I first started, one of the things that we actually used was the bedside table and the height of it was something that helped generate some of the flow, so that’s a long time ago. Knowing that, I have seen both of your publications. So, Dr. Ron Wald, in October 2023 in Intensive Care Medicine, you published as the lead author “Initiation of Continuous Renal Replacement Therapy Versus Intermittent Hemodialysis in Critically Ill Patients with Severe Acute Kidney Injury.” It’s a secondary analysis of the STARRT-AKI trial, correct?
Dr. Wald: That’s right.
Dr. Madden: Then, Dr. Koyner, just a few months later, in the Journal of Critical Care, you published “Initial Renal Replacement Therapy (RRT) Modality Associates With 90-Day Postdischarge RRT Dependence in Critically Ill AKI Survivors.” It’s pretty intriguing to me that two very different publications with two very different methods associated with it seem to come up with pretty similar findings. I’d love to hear each of you describe your study a little bit and the findings. Ron, I’d like to have you start.
Dr. Wald: Oh, for sure. Thank you. The concept of CRRT perhaps being associated with a lower risk of dialysis dependence after an episode of acute kidney injury is not a new concept. In fact, there have been several observational studies that were published on this topic in the past. Most of them are about a decade old. They were summarized in a meta-analysis by Dr. Antoine Schneider about 10 years ago. But since then, no one has really tackled the question, I don’t think.
So we saw an opportunity at the conclusion of the STARRT-AKI trial, which I was involved with, which looked really at the timing of dialysis initiation more so than what modality was used. But nonetheless, within the STARRT-AKI trial, there were 3000 patients recruited, of whom about 2100 initiated renal replacement therapy. So we said, well, even though the trial wasn’t necessarily about the topic of modality, CRRT versus intermittent hemodialysis, we nonetheless felt that, with all the data collected through the trial, we would be able to answer this question, recognizing that it wouldn’t be a randomized intervention of CRRT versus intermittent therapies, but rather an analysis that was really just a cohort study whereby the patients who were in the trial who received CRRT versus IHD based on clinician decision-making would be compared.
One can imagine that a typical patient who starts on CRRT is very different than someone who starts on intermittent hemodialysis. There are, as one would expect, patients on CRRT are quite sicker, require more organ support, and are generally individuals who may have a greater degree or a greater likelihood of dying as a result of all that, whereas those on intermittent hemodialysis may have a greater degree of hemodynamic stability, for example.
These can be obvious differences in the patients who are treated with each of these therapies because it’s clinicians who are making the decision. It’s not by random chance as it would be in an RCT. So in order to overcome the inherent differences between those who initiate CRRT versus those who initiate IHD, we developed a propensity score, in other words, the likelihood of a patient receiving CRRT based on all the variables that are in the database.
So we used that propensity score now, for the likelihood of receiving CRRT, to do what’s called weighting. What we did was, we said, can we balance out the two groups, those who initiate CRRT and those who initiate IHD, by taking this propensity score and giving different degrees of weight to the various patients in the cohort to create what’s called a pseudopopulation. A pseudopopulation means that we’re not necessarily comparing one person who received CRRT to another who received IHD, but it could be the population when kind of inflated or deflated by these weights.
Ultimately what’s created is a balance of baseline characteristics between the two groups. So, once we were able to create this kind of balance between the baseline covariates between those who initiated CRRT and those who initiated IHD, we then felt comfortable, okay, now we can compare the two. Our primary outcome was a composite of death or dialysis dependence by 90 days.
We showed that those who initiate CRRT were less likely to die or receive or be dependent on dialysis if they were alive at 90 days. But that most of that difference, most of that quote unquote benefit attributed to CRRT was driven by a much lower degree of dialysis dependence at 90 days. So our summary would be that it’s not clear that CRRT saves lives or reduces certain mortality and, as has been shown in multiple trials in the past, but that CRRT initiation associates with a lower degree of dialysis dependence, greater degree of renal survival, at 90 days. If a patient is lucky enough to survive to day 90 of their critical illness, CRRT seems to associate with a higher likelihood of being dialysis independent.
Dr. Madden: Jay, would you mind then giving us an overview of your study?
Dr. Koyner: For sure. There certainly are some similarities, but we did not start with a randomized controlled trial. We started with a retrospective database and we used the premier health database, which is a hospital database here in the United States that accounts for maybe about one in five of every hospitalization in the United States and is a representative sample of hospitals in the United States with regard to hospital location, with regard to type of hospital. When I say location, I just don’t mean geographically in terms of the west, the east, the south, the midwest, or wherever but also urban versus rural, as well as the size and nature of the hospital, teaching versus nonteaching hospital.
What we did is we queried that database to look at patients who received dialysis as part of a diagnosis of acute kidney injury. Then we wanted to specifically look at patients who had at least 90 days’ survival after they acquired this dialysis-requiring AKI.
Then, similar to Ron, we wound up obviously having to look at patients who had CRRT and IHD. They are different, and we demonstrated that there were some of the same differences that Ron described, that sometimes patients who require CRRT are both younger and perhaps, you can’t see me, but I’m using air quotes, sicker, whether that’s measured as part of a risk score or as measured by a need for mechanical ventilation and/or need for vasoactive meds.
We attempted to match people using inverse probability treatment weighting, similar to the idea that Ron discussed, where we tried to make sure that the patients who received IHD as their first modality looked as much as possible like those who received CRRT as their first modality. Then we followed them out and we attempted to see whether or not a) they lived longer, but b) whether or not the modality choice impacted their kidney recovery at 90 days. We did this in a variety of ways.
As you may know, it is difficult sometimes when someone leaves the hospital to know if they continue dialysis. The definition of being dialysis dependent is oftentimes very variable in the literature. So we used a variety of definitions, whether or not that was one dialysis session within the last three days or two in the last five to seven days.
Either way, we were able to demonstrate that the patients who received CRRT as their first modality were more likely to have renal recovery. We did this in a variety of subsequent analyses where we did things like exclude, using both of those definitions of dialysis dependence, but then also excluding patients who come from hospitals who only had IHD as an option, as well as excluding folks who we did not have claims data for, by which I mean we tried to link this data with hospital claims data to try to make sure that all of these people were still alive and actively getting care 90 days out.
Dr. Madden: Thank you. Having read both studies, we keep talking about 90 days, but I had noted in Ron’s study they were talking about 90 days after randomization, whereas Jay, in your study, you’re talking 90 days post-discharge. When I look at some of the data in regard to the patient population for Ron, some of the patients spent approximately 40 days at the upper limit in the hospital. How does that have any correlation?
Dr. Wald: Well, there’s a certain degree of arbitrariness to when you look at whether a patient is dialysis dependent. The traditional choice of 90 days comes from the fact that in the world of nephrology, where we spend a lot of time with patients who have end-stage kidney disease or kidney failure, the transition between acute kidney injury and chronic kidney disease is traditionally defined at 90 days. In other words, if a patient has a certain degree of kidney function at day zero, you assess their status at 90 days thereafter to see if their kidney function is worse than it was at day zero. At day 90, usually we refer to it as chronic kidney disease. So day 90 is kind of this traditional time point for defining chronicity or irreversibility.
That’s why we chose that time point. It was also convenient for us in our trial, because that’s where we collected the primary outcome for our trial and don’t have as much data past the day-90 point. In fairness, if I was interested in seeing whether a patient was dialysis dependent, I would want to go as far to the future as possible, kind of the way Jay did it, but we were somewhat limited by the data collection in our trial.
Dr. Madden: So there are some study limitations in that respect, as we look at it. Jay, any comments? It sounds like the 90 days, as it was just explained, is the benchmark.
Dr. Koyner: I think 1000%. I think unfortunately for you, you’re on a podcast with two nephrologists and even more specifically for me as an American nephrologist. It is oftentimes at 90 days, once someone has been on dialysis for 90 days, that they then become eligible for some of the government benefits, or at least in the past, it’s been the case. So for many studies looking at dialysis dependence, that 90 days becomes a key marker.
That said, we know that there are people who recover their renal function after an AKI after 90 days, but it is one of these things that has been translated, perhaps incorrectly, but in this case, usefully from CKD to AKI care, that 90-day mark is usually when there are some transitions with regard to reimbursement in the United States, as well as around some of the care.
Dr. Madden: You just referred to the fact that as a U.S.-based physician. For the STARRT-AKI trial, though, there were 15 countries and 168 sites. So there is some degree of difference as we talk about the healthcare systems, potentially how the individual renal replacement modality is implemented and run. So from both of these publications you have, what does the clinical evidence show regarding the initial mode of therapy and dialysis dependence at hospital discharge?
Dr. Wald: I think that ideally when we translate the medical literature into clinical practice, we like to have the highest-level evidence whereby patients are randomized to two different treatment strategies, and the one that yields the better result is the one that carries the day and ultimately should translate into clinical care, assuming that the user of that knowledge feels that the trial was conducted in a way that’s similar to how he or she might take care of his or her own patients.
But I think that, in this particular case, we have to remember that these are observational studies, meaning that, despite all the statistical rigor that we put into these studies to create some balance between the CRRT and IHD initiators, inevitably there will be what’s called residual confounding, things that we can’t account for when you compare CRRT and IHD.
So, while the results are hypothesis generating, I would be careful to say, well, CRRT definitely protects kidneys and reduces dialysis dependence. I believe that there’s a lot of good reasons for that to be true. And I still am a firm believer in using CRRT, particularly in the most vulnerable patients who are unstable and for whom intermittent hemodialysis is an inappropriate choice. But I still want to be cautious when interpreting or inferring knowledge from studies that are not randomized, such as these.
Dr. Madden: Thank you. Jay, do you have any comments you’d like to make regarding that?
Dr. Koyner: I think that Ron is correct as it relates to both of our studies. Neither one of them seems ideally suited to definitively answer the question. But I also feel that there is mounting evidence that neither one of our studies is the first to perhaps demonstrate that there is potential benefit to doing CRRT, recognizing that one of the goals in some patients is to be able to do the role of dialysis in someone who is hemodynamically unstable.
That said, I think that he’s right. Even though we both demonstrate that there is less dialysis dependence in those who receive CRRT as their first modality, it’s a lot more complicated than that, right? That in many parts of the world, CRRT is run by nephrologists. In many parts of the world, it’s run by intensivists. In some parts of the world and in some countries, it just depends on where you are in that country in terms of who’s doing it. All of those things probably or could potentially play a role in terms of how it’s delivered and when it’s delivered.
And we’re sort of lumping all together, if I’m being honest with you, all of CRRT as one thing when we know that there are multiple modalities within CRRT. I think that that’s a whole other podcast. I know that both Ron and I have also published data looking at different modalities and perhaps their impact on patient outcomes. There too, the data doesn’t necessarily show that there’s a benefit and that maybe it is just, hey, all CRRT may give you a better chance of having renal recovery, even in those who are fortunate enough to survive their critical illness or their AKI.
I think that there’s what I would describe as cautious optimism around the idea of continuing to use CRRT as the preferred modality rather than definitively saying you have to do it because I don’t know that we have the data that supports that.
Dr. Madden: With that said, and we’re talking about renal function and dependence upon renal replacement therapy, so IHD, once they’re out of hospital, is there clinical evidence that’s demonstrating a mortality benefit in which initial treatment with CRRT versus IHD with AKI?
Dr. Wald: I think that the trials that have been done to date, and they haven’t been very large, but those that have been done and have been summarized in meta-analyses show that CRRT does not confer a survival advantage over intermittent hemodialysis. And I think that having looked after patients in the ICU for over two decades, my view is that the patient’s underlying illness, the reason that brought them to the ICU, often the severe sepsis or some other complication of surgery, that’s usually the factor that will determine whether that patient is fortunate enough to survive or not. It’s unlikely that one single aspect of their care, when their care is so multifaceted, namely whether they get CRRT or IHD, will determine whether they survive or not.
I think the trials were designed to show a mortality benefit of CRRT or to test whether there was a mortality benefit with CRRT. It’s unrealistic to think that one relatively small component of their care could do so much. But nonetheless, we have to think, in my view, a lot about how patients survive their critical illness and if, once they’re out of the acute phase of their illness, we want them to be as healthy as possible, return to their baseline function as possible, and return to their baseline organ function as soon as possible.
We know that, for example, the way we provide mechanical ventilation will have an impact on the patient’s lung structure and function at the end of an ICU stay. There’s a lot of evidence around that. I think by the same token, we have to appreciate the fact that the way we deliver renal replacement therapy does have implications. It may not make the difference between life and death in the ICU, but it may certainly have an effect on the ischemic burden that we place on the already injured kidneys.
I therefore believe that it would be wrong to, as some people did, to say, well, CRRT doesn’t yield any mortality benefit, and it’s more expensive, and it’s more resource intense, so let’s abandon it and just do intermittent hemodialysis on everyone. I think that’s wrong. Our studies and those that preceded us suggest that indeed, where CRRT may not save lives, it may actually prevent people from being chronically dialysis dependent.
As someone who looks after patients with kidney failure who receive maintenance dialysis, that’s a very, very big intrusion into someone’s life that in and of itself downstream is associated with higher risk of death. I would love for all the patients that I look after in the ICU, even if they are lucky enough to survive, I’d like them to be off dialysis and ideally with a level of kidney function that is as close as possible to the kidney function they had before they got sick.
Dr. Madden: That’s a great hope. I think that’s what we all strive for when we’re taking care of our patients in the ICU so that we can optimize their outcomes and really have quality of life. Jay, did you have any comments you wanted to make?
Dr. Koyner: Again, I’ll echo what Ron said, recognizing that in our study, we specifically were only looking at survivors of the hospitalization so that we took the CRRT-versus-IHD mortality benefit out of the equation because we were already looking at that selected population that you heard Ron talk about in the very beginning. I’d also just argue it’s not just in my mind about the patient’s kidneys.
I have to believe that 20-plus years of working in ICUs has taught me that if you’re trying to keep someone even net even with IHD versus CRRT and they’re getting, on average, let’s say two liters or two to three liters, that it’s not just doing the work of the kidneys that the dialysis is doing. But if you’re doing that over 24 hours as opposed to over three to five hours, let’s say, with an IHD treatment, the strain on their heart and their cardiovascular system is going to be much less.
I think that we as nephrologists sometimes underestimate the strain that we put on the rest of the patient’s body by putting them on a dialysis machine and pulling two, three, four, five-plus liters in a couple of hours. And now that only gets amplified in a critically ill patient who’s on vasoactives or maybe even had their heart operated on a few days ago.
I think that it’s nice and I am with Ron that I want all my patients to return as close to their baseline kidney function as they can. But it’s also about not doing harm to other organs, which I think that there’s the potential for, even though we may not have the ideal tools to identify who’s going to survive and who isn’t.
Dr. Madden: I have to say, I’m sorry, but we’re really out of time. Before I close out this podcast, I’d like to ask Ron and Jay if either of you wanted to bring up that we didn’t have the opportunity to touch on yet.
Dr. Wald: I want to echo Jay’s last point with regard to fluid removal. I think that beyond the benefits of CRRT, particularly in the hemodynamically unstable patient, irrespective of hemodynamic instability, in patients who have dramatic amounts of fluid overload, CRRT is probably the most reliable and effective way to remove fluid. So we shouldn’t just reserve it, so to speak, for patients who are on vasopressors.
Dr. Koyner: I entirely agree. I do want to say, I think that I misspoke when I talked about some of our secondary analyses. I think that we looked at people who had received one RRT treatment in the last three days, and then it’s greater than two or greater treatments in eight days. I think I said five to seven days, but it’s eight days, but I don’t have too much more to add.
Dr. Madden: All right. Well, thank you for that. This concludes another episode of the Society of Critical Care Medicine podcast. If you’re listening on your favorite podcast app and you liked what you heard, consider rating and leaving a review. For the Society of Critical Care Medicine podcast, I’m Maureen Madden.
Announcer: This podcast is sponsored by Vantive U.S. Healthcare. Vantive supports true patient-focused treatments with industry-leading CRRT technology and is a partner dedicated to optimizing your clinical success in treating patients with acute kidney injury. Our commitment to you starts with education and provides complete support every step of the way. Visit us at vantive.com.
Maureen A. Madden, DNP, RN, CPNC, AC, CCRN, FCCM, is a professor of pediatrics at Rutgers Robert Wood Johnson Medical School and a pediatric critical care nurse practitioner in the pediatric intensive care unit at Bristol-Myers Squibb Children’s Hospital in New Brunswick, New Jersey.
Join or renew your membership with SCCM, the only multiprofessional society dedicated exclusively to the advancement of critical care. Contact a customer service representative at +1 847 827-6888 or visit sccm.org/membership for more information. The SCCM podcast is the copyrighted material of the Society of Critical Care Medicine and all rights are reserved.
Find more episodes at sccm.org/podcast. This podcast is for educational purposes only. The material presented is intended to represent an approach, view, statement, or opinion of the presenter that may be helpful to others. The views and opinions expressed herein are those of the presenters and do not necessarily reflect the opinions or views of SCCM. SCCM does not recommend or endorse any specific test, physician, product, procedure, opinion, or other information that may be mentioned.
