Bharath Kumar: Hello everyone, welcome back to Crit-No. I'm Bharath Kumar and I'm joined by my co-host, Dr. Ramesh Venkatraman. Before we get started with today's episode, as always a reminder that if you've been enjoying our content, please comment, like, subscribe, share, and spread the word. Thank you so much again for all of the feedback and the support. Today we are again focusing on one of those topics or trials that has a lot of relevance to our everyday critical care practice. We encounter patients with acidemia pretty much every day in our intensive care unit and there's always been a debate as to whether we should try and correct this with infusions of sodium bicarbonate. There have also been some trials in the past and today we're going to focus on one of the recent trials, the Soda Big trial that evaluated the usefulness of sodium bicarbonate. But before we get into the trial itself, I'm going to request Ramesh to... talk a little bit about the rationale for whether we should use sodium bicarbonate and whether we should be evaluating this as an interventional strategy. Over to you Ramesh.
Ramesh: Thanks Bharath. A lot of good trials recently so it was hard kind of for us to pick but we thought this is something that we do almost every day in the ICU that it would probably top the list of the trials that we would like to do and we'll probably get to more interesting trials in the near future. So going on to this trial, the first fundamental question is as a clinician, why do we worry about acidity? And all of us know and that's pretty clear that respiratory acidemia behaves differently than metabolic acidemia. The impact of respiratory acidemia is different than metabolic acidemia. And most of us worry about the metabolic acidemia much more than respiratory acidemia. And the reasons being, the common physiological reasons being one, yes, we need to treat the cause of ongoing reason for the metabolic acid generation, number one. But the acidemia itself, one, demands a higher physiological reserve. To compensate, you're going to stress your physiological reserve even more, commonly by hyperventilating. especially if your kidneys are impaired, then a bunch of load is just taken by the respiratory system. So one, you may need to escalate the level of support. You need to provide someone to be able to augment their physiological reserve, number one. Number two, acidemia directly or indirectly potentially causes more organ dysfunction. The two commonest doubted mechanisms are one, it decreases the catechol response when you especially when you have vasopregia. So, when we use catecholamines to increase the mean arterial pressure, our responsiveness may not be as good and we may need a much higher doses of vasopressor which has been clearly shown to cause more ischemia and more. more deleterious effects. Second thing is, acidemia itself can impair myocardial dysfunction. So between not responding to your vasopressors and having myocardial dysfunction with the underlying reason for metabolic acidosis, directly or indirectly are going to propagate more and more organ dysfunction. And at some point we need to halt this. And we also need time to correct the underlying cause of the metabolic acidemia and we need some kind of a temporizing measure that would help us get to the next few hours, if you will. So those are generally been the rational about why we need to use bicarbonate in the ICU. And there have been a couple of studies that have addressed this question. So I think it's probably worthwhile addressing those two studies first and then talk about why the study evolved from those two and then go on from there. What do you think?
Bharath Kumar: Yeah, absolutely. mean, I think what you're describing is sort of a double or a triple whammy, right? The metabolic acidemia in particular is placing additional demands on already stretched reservoir system. And in addition, it is also blunting whatever we are trying to do to make things better. So it's sort of a combination of multiple mechanisms at play. So the first trial that looked at this particular or the first big trial, would say, we are only covering the major trials that have looked at this issue is the Bicar ICU trial that was published in 2018 in Lancet. This was from a French group conducted across 26 ICUs in France. They essentially enrolled patients, adult patients in the ICU that had severe acidemia, pH of less than 7.2, bicarbonate of less than 20 millimoles per liter, a PACO2 less than 45.
Ramesh: .
Bharath Kumar: a sofa score of four or more and a lactate of two or more. So that's a a pretty sick patient cohort in terms of or moderate to severe patient sickness cohort with severe acidemia. They specifically excluded patients with respiratory acidosis if there was proven digestive or urinary tract losses of bicarbonate, ketoacidosis, stage 4 CKD and so on. Their intervention was, and I'm specifically saying this because the intervention is different in different trials, they used a 4.2 % solution of sodium bicarbonate. And their goal was to achieve a pH of 7.3 or higher. And they continued these infusions until day 28 as and as needed. Most patients, of course, didn't need it for longer than 36 hours. But their intention was to continue it for until 28 days. The comparator arm was standard care in the Bicar ICU trial. They had standardized the indications for renal replacement therapy, both at admission as well as at 24 hours into the patient's stay. And their primary outcome was a composite outcome of death by 28 days or at least single organ failure, presence of a single organ failure at day seven. And there were multiple secondary outcomes that they looked at. So very quickly, they included 389 odd patients. Most of these patients were on ventilator or vasopressure. 80 % were on ventilator or vasopressure. And as I said, achieved a pH of 7.3 and maintained this for 36 hours or more, although the intervention itself could be given for longer. There was no difference in the primary outcome in this trial, in that composite outcome. There was 71 % in the control arm and 66 % in the intervention arm. This was not statistically significant. The only important sort of, there was a pre-specified subgroup. They had patients with AKI stage two or three as a pre-specified subgroup and they looked at their effect of the intervention. And there, there seemed to be a benefit in that pre-specified subgroup. But there were only about less than 50 % of the patients where patients at stage two or three AKI. So that's the Bicard ICU. And of course, patients overall also needed less RRT in the bicarbonate arm and had more RRT-free days. So this was the first trial looking at this issue. So overall, no signal. But it seemed like in the Stage two or three, AKA, there may be a benefit. Yeah.
Ramesh: No, I think that's fair because I think pathophysiologically the metabolic acidemia that occurs from AKI, I think is phenotypically different from metabolic acidemia occurring from maybe other causes or let me put it this way metabolic acidemia of any cause associated with an AKI. behaves differently because what we talked about the physiological reserve is even more diminished and there is probably more urgency and more need for us to correct the acidemia exogenously when your kidneys cannot. So, I would say it makes physiological sense that that subgroup may benefit more compared to. people who don't have the AKI. As you rightly pointed out the study, overall outcomes were equal between the two arms and only less than 50 % had AKI. A couple of points worth mentioning here Bharat is they use bicarb in a completely different way than what I would use. You did point out the 4.2 % but they also used it in a way that I normally wouldn't use it. So the 4.2 % is different than the 8.4 % that we use in India. So just for the audience to know. And they also created a solution of 125 to 250 ml volume, not the traditional 500 ml volume we use. And they gave it over five hours. And then they could repeat up to four doses, which is a liter in any particular given day, till the pH goal was corrected. So that is probably not how you and me use bicarbonate in the ICU. So that's. Now, one thing that I would like to point out to the audience here that slightly different how they used it.
Bharath Kumar: Yeah, yeah, different. And we have, I think, access to 7.5 and 8.4%.
Ramesh: Yeah, yeah. So then they go on to the BiCAP ICU 2 trial where I think the same French group about 43 ICUs in France. This time they basically said let's look at the subgroup with only severe AKI, moderate to severe AKI because that's where the signal was in the first study. So they basically again an open label study. Again this time they were stratified only for age more than 65 and extreme acidemia, a pH of less than 7.10 or a pH of more than 7.10. The inclusion criteria was pretty similar your pH has to be 7.20 or less, your bicarb has to be 20 or less and your PaCO2 should be less than 45 because you do want respiratory acidemia in that. This time the outcome they looked at was a 90 day mortality and they wanted a 10 percent reduction considering that the bicarb 1 ICU study had 70 percent mortality but that is the overall mortality. But in AKI you would expect more. They said, OK, my 90-day mortality is going to be 80 % and can I drop it to 70 %? And that's the power analysis they did in this particular study. They included patients with KDIGO class 2 or 3, so that that's moderate or severe AKI. The interesting thing here, Bharat, is they almost screened 3,000 patients and ultimately randomized only 640, which means that 20 % of the patients, so 80 % by far, were excluded. So 320 per arm. Again, 55 % at septic shock, at the arms mechanical ventilation again was close to 80%, vasopressors close to 80%. The baseline pH was quite bad, 7.15, which all of us would react to. Lactate level close to six and a creatinine of 2.3. So these are pretty sick patients that you and me would see and I think probably would start using bicarb in these situations if we ever get to it. They again use same 4.2 percent bicarbonate, the same dilution of 125 to 250 ml and they gave it over 30 minutes the solution and then they could repeat it up to 1 litre within a particular day till the pH reach 7.13 or higher in both the groups. And again like the bicarb once trial they basically standardize the indications of initiation of RRT between the two groups. What the study showed is there was no difference in the 90 day mortality. The control arm and the bicarb arm had about 61 and 62 % mortality. Again, the use of renal replacement therapy was lower in the bicarb group. It was only 35 % versus 50 % in the control group. The other thing we talked about is catechol responsiveness. The use of vasopressors was similar, almost 94 % in one group and 97 in the other group. And time to initiation, they were also able to postpone dialysis in the bicarb group. by about 15 hours. That 31 hours is the time to initiation in the bicarb group and only 16 hours in the control group. So I think these two put together tell us that bicarb infusion helps us probably postpone renal replacement therapy and maybe avoid renal replacement therapy in a small proportion of patients. But it really doesn't improve our vasopressor requirement or mortality based on what we know from these two studies. The downside of these two studies, Bharath, apart from the bicarb is that they were both openly labeled. The clinicians knew who got bicarb and who didn't get bicarb. And we all know that the decision to initiate RRT is quite subjective. At this point of time, there is no right timing for us to start RRT. And it's kind of a clinical gestalt. that decides who gets RRT. And in an open labeled study and looking at RRT as one of the outcomes, there is a possibility that there could be some kind of a bias that could have affected the outcomes. So those were the initial thoughts that went through. And then I think the current study wanted to go one over and see if we can take the good things out of the previous study, try to see if we can weed away. some of the limitations of the previous study and design a new study. think that's what happened at the Soda Big Trial.
Bharath Kumar: Yeah, I think we have reasonably covered the background and the context leading into this trial. Yeah, so with the Soda Big Trial, as you rightly pointed out, they wanted to choose, I think, relatively less sicker cohort and get the patients in a bit earlier, I think, with a pH when it's less than 7.3 itself.
Ramesh: Thanks for listening. Actually, I just want to interrupt you just for a second. I think that's a very valid point you make. I think this is something that I wanted to bring out. think now that you said it kind of prompted me to say that. When we talk about all metabolic acidemia, not all metabolic acidemia is going to either hurt us, not that all metabolic acidemia is going to kill us also, right? There is some amount of metabolic acidemia that happens from purely from a stress response. And once the acute stress response is taken away, the acidosis results spontaneously.
Bharath Kumar: Sure.
Ramesh: So, that group is probably not going to benefit from a bicarbonate therapy. There is also a group which basically has uncontrolled acid production from whatever bad etiology, let say an ischemic gut or a ischemic hepatitis, where the acid production is just relentless and you just trying to neutralize it with bicarb is really probably not going to change outcome. And the other thing you could argue is if I allow the pH to go below 7.2 before I start bicarbonate. most of the damage is already done, right? We talked about other organ dysfunctions and all that. So I think all of that kind of factored in and they said, okay, is there a role in me trying to start bike up a little bit early? I think most of us would agree 7.25 is probably somewhere we would consider, 7.2 definitely, but one way of looking at it is 7.2 maybe too late. By then there's a lot of damage already done to the other organs and maybe I should have started it early. So I think... the 7.3 cut off is probably derived from this kind of a thought process.
Bharath Kumar: Yeah, absolutely. So they were trying to, as you said, maybe plug in or fix some of the concerns from the previous trial. At least that, think, was the intention. So the research question in the Soda Big Trial, so we'll get to the Soda Big Trial now, was among patients, adult patients admitted to an intensive care unit on vasopressors, does an infusion of sodium bicarbonate lead to a lower risk of make or major acute kidney events at day 30? So a different outcome than the other trials, the primary outcome at least. So this trial enrolled in 55 ICUs in seven countries. But the maximum patients, actually more than 90 % were from Australia and New Zealand, only a handful of patients from the other countries. The population they enrolled was adult patients receiving a continuous infusion of vasopressors and with metabolic acidosis, a pH less than 7.3 as you mentioned. Base excess not more than minus four, a PaCO2 not greater than 45. If they were intubated, it was 50. Again, like the previous trials, they excluded patients that had DKA, those that had major bicarbonate losses from the GI or urinary tract. They excluded patients that had a high risk of cerebral edema, those that were at risk for imminent death, those that are already on RRT at the time of randomization or planned for RRT imminently, or if they had a EGFR less than 30 due to CKD. So this is the population that they wanted to include. The intervention here is 8.4 bicarbonate, which is diluted in 500 ml of 5 % extras and given over five hours. So they also had a protocol in this particular trial on how the infusion would be administered, including some rules for titration, up titration, and down titration, really down titration. So if the pH was less than 7.3, they would continue to keep the infusion at 100 ml per hour. If it is between 7.3 to 7.3, they would bring it down to 25 ml per hour. It was a 7.35 they would stop the infusion. But the entire duration of the infusion itself is only five hours. In the control arm, patients would be given 5 % extras. Both the infusions were prepared by somebody that had nothing to do with the trial, an independent person preparing both the infusions. They obviously look alike, so they just delivered to be hung. And after the five hours, the use of bicarbonate or not was entirely left to the. treating teams or the treating clinicians. So really, the study is testing only a five-hour infusion once these criteria are met. And all of these criteria had to be met within the two hours prior to randomization. And once they are met, they obviously want to get the infusion going as quickly as possible. The outcome they chose is called as the Make 30. intensive care clinicians are going to be familiar with this outcome from some previous trials. It's a composite of death, all-cause mortality, receipt of renal replacement therapy, or persistent renal dysfunction, which was defined as a creatinine value, which was at day 30, more than 200 % of baseline. And they were assessing all of these outcomes at day 30. They had multiple secondary outcomes. The trial used a 1-1 randomization. They stratified by site, a pH level below or above 7.25, a creatinine level below or above 1.7. I'll also quickly talk a little bit about the sample size assumptions. They assumed a 40 % event rate in the control arm, 40 % event rate of the make 30 outcome. And the assumptions that drove this came from the Bikara ICO trial. And they wanted to detect a 14 % difference in that primary outcome, again based on Bicar ICU data. And they say that this is a slightly conservative assumption as compared to that. This gave them a sample size of 500. I would say broadly, very appropriate analytical techniques. The one interesting thing that they did do is that they had something called a blinded sample size re-estimation after the first 100 patients. to see if they would need to increase their sample size. So it was a bit of an adaptive trial in that sense. So this is broadly the peak of the Soda Big trial.
Ramesh: Nice one Bharath. Bharath, I just want to point to the audience this the word adaptive here shouldn't be confused with adaptive platform trials. Right? I mean that's a common most people are now starting to hear the term adaptive platform trials. So this adaptive is not the same as that. So could you explain the word adaptive in this particular trial and you talked about blinded reassessment of the sample size. So can you just elaborate on those two so that we kind of know why you do that and what the impact is on the study?
Bharath Kumar: Sure, sure. Yeah, I mean, adaptive platform, would say, is sort of within the family of adaptive, but not all adaptive trials are adaptive platform trials. So when you're talking about adaptive, what we mean is that we are, adaptive trials are those trials that offer the flexibility of modifying certain elements of your protocol. but without compromising the integrity of the trial. So after the trial has started, adaptive trials give you the flexibility of changing certain elements. They obviously have to be pre-specified what you intend to change, but you do it in a way that it doesn't compromise the integrity of the trial. Now, number of things can be adapted. So you can add or drop arms, for example. That could be a form of adaptation. Adaptive platform trials typically do that. They have multiple domains, multiple interventions. They add, they remove, and so on and so forth. They can get as complicated as it can. And then you can also do simple things such as change the, so you may set in certain rules to increase the sample size if needed. So that's what these investigators have done. So why would we do that? As I said, when you calculate your sample size, you rely on some assumptions which are based on previous data. And you also make assumptions about what your effect estimate might be. Now, these assumptions could be wrong, because in this case, they are looking at the Vicar ICU, happened eight years back, or seven years back, or six years back when they started. And they're doing the trial in a different country, different environment, all of that. So the assumptions that are there could be wrong. So what you could propose to do is, which has to be pre-specified as as these investigators did is once 100 patients were enrolled, they had an independent team that looked in a blinded manner at the overall event rates. And using those event rates, they re-estimated their sample size to see if the assumptions have changed, does the sample size change dramatically as well. But what they did importantly here is they capped what the new sample size can be. Because you know, for example, if your assumptions are way off the mark, but the new sample size comes back as 2000 or 3000, right? They were only powered. mean, they initially started with 500. Their budget, their resources are all going to be for 500. Now, if it comes up as 2000, they can't do that trial, right? Even if that's the number comes up as. So intelligently, they said we are willing to go up, but we are capping that at 700. Now it gives them the flexibility to say, if you get our assumptions wrong, we are willing to go up to 700, but not anything beyond. Now it can also be the other way around. Your assumptions might start working. So instead of the numbers that you estimated, the numbers may be much higher. The control arm, they know that. In this case, potentially one might argue, can you reduce the sample size? You could, but that would always be looked at with a lot of suspicion when you try to drop your sample size. So typically, investigators don't try to drop the sample size. they may increase it but with a certain cap. So that's something that these investigators did. So this is the kind of adaptation they did. One last point, another form of adaptation that could also be done. There are other types, but I'm just talking about one or two examples here is you could also change your randomization ratio. So you can do something as an adaptive randomization ratio where as the trial progresses, you might give the computer an instruction to say slightly increase the ratio in favor of whichever arm is doing better. Again, it's blinded. The investigator would not know this. And the increase would be very marginal, like 1 is to 1.05 or 1 is to 1.1. or something like that. So you sort of, in some ways it's called playing the winner, right? But you bias it slightly in favor of the arm that is doing better at every interim analysis, right? So that is something that could also be done. So there are different ways of adapting. In this case, they have done one particular adaptation.
Ramesh: fascinating, right? I mean, this is a new concept for all of us. And I think in every one of our podcasts, we are trying to elaborate on one statistical or one methodological nuance that I think we learn from it. And therefore, hopefully the audience also enjoy and learn it. this is nice, adapting your sample size based on how you go. And they found that they didn't need to do anything. They stuck with their sample size of 500.
Bharath Kumar: Bring on ya.
Ramesh: To summarize, think they did all the things that the previous trials probably we thought of for one. They didn't pick two sick patients. They started out with a pH of 7.3 or less, rather than 7.2. The base excess is not terrible. They ruled out respiratory acidosis. And what we figured out that is that vasopressor responsiveness is something that we care about. And therefore, they said we need all patients to be on vasopressors. Remember that the previous two trials had only about 80 percent vasopressor rate here everybody's vasopressors we are hopefully seeing if that would be impacted but with bicarbonate. So I think they have plugged in a lot of the holes the previous study did and most importantly this is a blinded trial. So the last study last two studies were not blinded and that was one of the limitations and this one they just completely blinded. So they have bettered it based on the knowledge of the previous two trials and looked at relevant outcomes. So going to the results If you look at that, like you said, 500 patients, Bharat, age is about 65 and 68. Apache 2 score is about 21, 22. So reasonably sick patients, the baseline pH was 7.26. So right around the cutoff where you and me would start thinking of doing bicarb, so which is good. The base excess was minus nine, roughly in both the groups. So reasonably, you know, strict acidemia. creatinine was close to 1.4 and again 45 % of patients had acute kidney injury in both groups. So I think this kind of represents the patient population that you and me would see and probably start bicarbonate therapy if we think that would help that particular patient. The other thing they did is they used 8.4 % bicarbonate, which is what we would use. But like you said, they used it only for five hours and then they kind of let the clinicians do. whatever they wanted to do. However, if somebody has a pH of 7.10 or less, they could go to open label bicarb. They had certain kind of safety mechanisms, bicarb less than eight, base deficit minus 15, potassium of more than seven, or a pH less than 7.1, you could go to bicarb. Or you could do anything you want after the five-hour study period is over. And if you look at the make outcome at 30 days, make again, like you said, it's death. needing dialysis and renal dysfunction put together, was no difference between the two groups at 30 days. One group had 40.2, other group had 39.4. Among the multiple secondary outcomes, I just want to mention a few of this. one death at 30 days, no difference, 25 percent versus 24 persons. The need for renal replacement therapy within 30 days. lower in the bicarb group again consistent with the previous two bicarb studies. So 16.8 versus 20.9 there is a 4 percent absolute reduction in the need for bicarbonate. RRT dependence at 90 days again lower in the bicarb group 5.7 versus 9.1 so again roughly about a three and a half to four percent difference and renal dysfunction at 30 days again a four percent difference 14 versus 18 percent. bicarb again no major outcome benefit based on the primary outcome, but signal towards maybe some kind of minimizing the need for renal replacement therapy or probably no dependence on the replacement therapy at 90 days is something that they have looked at. And again, they included AKI stages two and three. think overall, I think this was kind of a population that Both of us would agree needed to be evaluated. So they were right on in terms of picking the right patients that you and me care about in our ICU. They gave bicarb in a way that you and me would probably give. One thing I learned in this particular study is to use something like a titration protocol. We generally don't do that in the ICU. We arbitrarily start at 50, 75, or 100. Here, they nicely structured an algorithmic approach to titration of bicarb. sort of bicarbonate. So if and when I use sodium bicarbonate, I think this is a great way to do it because now you probably minimize fluid overload not necessarily giving bicarb much more than I need to at the same time I would be aggressive enough to get the pH above 7.3 which is probably what most of us would say. So one big lesson for me before we go on to looking at the study in detail is if I'm going to use bicarbonate maybe something like this a titration protocol and maybe adapting their protocol in our ICU may not be a bad idea. see if it works for us and then maybe incorporate it to minimize the amount of bicarb we give. Mainly, I'm more concerned about the volume. Most of us start at 100 ml and sometimes it's on for six hours or eight hours and that 800 ml gone in somebody who's oligoanuric, right? So that's one point that struck me. And their recurrence rate of metabolic acidemia and the soda bicarb bomb was also not too bad. So the five-hour infusion seemed to correct it. Either you did better or you ended up on renal replacement therapy. So I think it's a...
Bharath Kumar: Mm.
Ramesh: If I have a titrated trial, it's not a bad way to use bicarbonate is one of the messages I would probably take home from this particular trial. What do you think?
Bharath Kumar: Yeah, I know. I agree with all of those important considerations, especially around the protocolization, avoiding volume or excess volume. I did think that this was overall, obviously, coming from a group with lots of experience, lots of credibility. So robust methods, I would say, overall, as you said, they tried to enrich the trial population by including patients that are on vasopressors and starting the infusion as early as they could. Most of these patients really get it within half an hour or 40 minutes of randomization. So I think they did all of that very well and the protocolization as well. I'm going to be a bit of a devil's advocate to raise some concerns or limitations. So the first one that I had was, was there a strong rationale for this particular trial? Between Bicar 1 and Bicar 2, I thought personally that we had answered or reasonably answered this question. So we were talking about why they did this trial. One is the fact that it was open label. The previous trials were open label. So yes, they do blind here. But the concern that always comes is how effective is the blinding? Because clinicians, they're monitoring the pH at 1, 3, and 5 hours. So people are going to get to know to some extent what infusion is going on. And we know that it's not a lot, but 15 % of the placebo group crossed over and got some bicarbonate, open-label bicarbonate, as compared to 7 % in the intervention arm. So clearly, clinicians probably guessed that they were on a placebo at that stage. The second thing was they said we wanted to enrich by having vasopressors as a mandatory criteria. I think that's important. But Bicar 1 and Bicar 2 didn't have that as a big problem. Like 80 % were actually on vasopressors, and they had a sick cohort as well. The mortality in Bicar 1, I'm not wrong, is 60 % or 70%. Yeah, right. Whereas here, it's 25%, of course, different continents, different timelines. All of that is there.
Ramesh: 70 percent of the control.
Bharath Kumar: The next point that I want to bring up is, was the effect estimate that they chose very optimistic, a 14 % reduction. Yes, they retested their assumptions in that sample size estimation, but nevertheless, the initial assumptions were probably, if there is a difference, it's probably very small and they're not picking that up. You and I have discussed this several times, Ramesh, the problem with these composite outcomes. They are all not created equal. The components are not really equal. Obviously persistent renal dysfunction is not the same as dying or being on RRT. I mean, in this case, there isn't a difference. But if there were to be a difference, then you're also saying, OK, it's only the creatinine that's changed or not changed. So did you want to come talk about
Ramesh: No, I'm just looking at the data from what you said. So the make was not different, 40 % in one group and 39 % in the other group. But if you look at the death, it's 25 and 24. So let's just cancel them out, right? And look at the other two outcomes, right? Need for renal replacement therapy, 16.8 versus 20.9. So the bicarb group has a 4 % advantage. And renal dysfunction at 30 days defined by 200 % from the increase in baseline, 14 versus 18%. So you get a 4 % advantage in
Bharath Kumar: Mm.
Ramesh: persistent renal dysfunction, 4 % advantage in terms of need for dialysis at 30 days and dialysis dependence, if you end up on dialysis, dialysis dependence at day 90, again an advantage of 3 and a half percent. So, people who die, die Bharat, but people who live may get less dialysis, may be on dialysis less, less dependent on dialysis at long term. I do know if those things. matter in this kind of a study, is very difficult because obviously study is not power to look at that particular question when the composite outcomes like you and me talked about a lot clumped together. But I cannot ignore the fact that I can avoid dialysis in some patients not all who survive and I also probably can have some patients off dialysis by 90 days in terms of their dependence right. I think both to me are important. So I think all of those signals are something that I can't ignore at this point of time. And my other biggest thought, Bharath, for all the blinding and everything that we have talked about in terms of trying to improvise on the previous studies, I would have liked them to standardize the criteria for RRT initiation. Both BiCAR1 and BiCAR2 did that. They didn't blind the fluid, but they standardized the RRT initiation. I don't see a reason why they couldn't have done that here, right?
Bharath Kumar: Yeah.
Ramesh: certain criteria or replicate the by car criteria. it worked there in two trials, hopefully it would work or develop a consensus in Australia and New Zealand and say we will have for this trial, these are the RRT initiation with some kind of a safety mechanism for clinicians to overrule. So one, blinding is not truly blinding. Two, the RRT decisions were kind of left open. And despite that, we saw a 4 % reduction in the need for RRT in one group or the other. So those are also some of the points that I want to know make and initiating RRT is also a clinical decision and an outcome right in this study. So I do not know clinical decision can stem from the fact that if I am able to maintain a pH above a certain threshold then I may not initiate dialysis. So my clinical decision depends on my pH improvement which BiCub would do anyway right. versus my outcome in terms of needing RRT is not dependent on just improvement of my pH. There are other factors that go into, right? I really don't know how to separate these two. Bi-carb is very good. If you look at the separation, the pH was 7.31 versus 7.26 after that five hours of infusion. So I get past the 7.3 mark. And if I have an arbitrary cutoff of pH in my mind for initiation of RRT, I probably will get away not.
Bharath Kumar: That too. Yeah.
Ramesh: starting by RRT in the bicarb group. So I also think the decision to initiate RRT is both a clinical decision that can be confounded by the pH and an outcome that is not necessarily confounded by the pH that is altered by the bicarbonate. So I also have a struggle in terms of figuring out whether it is truly a benefit or is it just that I am telling myself though pH is now above 7.3 maybe I can wait a little bit longer, right? That's the other.
Bharath Kumar: In terms of that, yeah.
Ramesh: Other point that I as a clinician sometimes comes to my mind with all these bicarb trials.
Bharath Kumar: Yeah, and I think that ties in nicely to the fact that despite, for the benefit of our reader, listeners, even though the numbers are all lower in the intervention arm, none of these are statistically significant. The make is not different. And the individual components, numerically, the RRT, all of that is slightly lower in the bicarbonate arm. But if you look at the confidence intervals, they're obviously wider. But it does.
Ramesh: Absolutely. Not, not, yeah, yeah.
Bharath Kumar: bring up the point that the evidence seems to be consistent if you take the car one, bike two, and then this that if there is a signal that's really in that, know, RRT, you know, avoiding or delaying RRT is sort of where it seems to be. I did a sort of a very rough meta-analysis pooling all of this, the three trials, Ramesh. You know, obviously they're not the same populations. So, you know, you'll have to bear with me. just, you know, it's a sort of a very. rough meta-analysis. And the signal seems to be the same. When you pull all of this data, there is no mortality signal. There seems to be a signal in terms of the RRT, the decreased need for RRT, even when you pull all of the three trials. So to me, in terms of my takeaway, and I'll come to yours, I felt it's reasonable to use bicarbonate, especially in severely acidic patients, pH less than 7.2. and if they have acute kidney injury, if you take all of the data from the previous trials and this trial. So I thought clinically my practice would sort of be aligned towards that now. What was, what is your takeaway on?
Ramesh: I agree with you Bharath, if somebody has moderate to severe AKI and metabolic acidemia, that subgroup, I think if there are no other compelling reasons for me to initiate RRT, no hyperkalemia, no anuria, no fluid overload and I think I'm just, I still have some amount of urine output left, my potassium is manageable, I don't have fluid overload, then I'll certainly try bicarb infusion. One thing I would change my practice in terms of how I use the BiCAP rather than a fixed dose or an arbitrary titration, I would probably have a titration protocol mimicking what they did in this ICU so that I can be a little bit more scientific about it. And if my five hour trial doesn't really work and I may probably say this patient needs dialysis or if I get a compelling reason, I will just move on to dialysis. I still think. moderate to severe AKI with metabolic acidemia, pH less than 7.3 with no other compelling reasons, certainly worthwhile, no trying bicarbonate. But the hope that even if it's a small proportion of patients, even if it is not statistically significant, there may be some benefit for me to avoid dialysis, even though I am not sure based on the existing data. But on the downside, Bharat, there is no much harm, right? And consistently all the three trials have not shown major harm with bicarbonate.
Bharath Kumar: my community. Correct.
Ramesh: And even in this trial, the harm that came was only from hypokalemia, right? I mean, nothing, no significant hypocalcemia and none of the other stuff. So that's not something that we can't handle. So if I put all those things together in that specific population I told you, I think the risk benefit favors BiCAP trial and in the hope that I can maybe avoid renal replacement therapy in the subgroup of AKI. That's what I would take home.
Bharath Kumar: Yeah. Yeah. No, absolutely. think that's a perfect take home for us to also stop this episode. So thank you. Thank you, Ramesh. And thanks, of course, to everyone that's been listening to these episodes and giving us lots of feedback. As always, a reminder again to please share, comment, like, subscribe. Thank you so much. See you.
Ramesh: And I'm also going to say that if there are so many trials out there, we just pick things that we think that everybody would benefit. But if there are good trials that we are missing out and that's something that you guys think we should look at it more, we are happy to dissect it and give our perspective on it. Just add it to the comments whenever you watch our videos and we'll take a look at it. And if it's reasonable, we'll do them too.
Bharath Kumar: Absolutely. Thank you.
Ramesh: Yeah, thank you. Bye bye.