Antipsychotics for ICU delirium don`t work: Podcast with Tim Girard

In this week's GeriPal podcast we talk with Tim Girard, Plumonary Critical Care physician-researcher at the University of Pittsburgh about his study NEJM study of Haloperidol vs. Ziprasidone vs. Placebo for ICU delirium in critically ill patients.

When I was on service last month, the ICU team said to me, "We've just started treating Mr. X's delirium today.  We started him on haloperidol."


Haloperidol and other antipsychotics for ICU delirium are ubiquitous.  When I was on service last month, the ICU team said to me, "We've just started treating Mr. X's delirium today.  We started him on haloperidol."  In this well-done randomized double blind placebo controlled trial, Girard and his co-investigators detected no difference between the 3 drugs.  Nada.  Zip.

Eric solicited questions in advance via twitter and many are incorporated into the podcast.

Some questions we address:
-But wait, might antipsychotics work for agitated hyperactive delirium?
-Why Ziprasidone?
-That's a lot of haldoperidol!  No geriatrician would recommend that much, what gives?
-Can we generalize this outside of the ICU?
-So...what else should we use instead? WWTDIC (What Would They Do In Canada)?

Enjoy!
Alex

Link:
-NEJM study of Haloperidol vs. Ziprasidone vs. Placebo for ICU delirium 

Listen to GeriPal Podcasts on:

• iTunes 
• Google Play Music
• Soundcloud
• Stitcher 

You can also follow us on Youtube!

Transcript:

Eric: Welcome to the GeriPal podcast this is Eric Widera… Alex I think it's your turn.

Alex: Oh sorry I was reading song lyrics. This is Alex Smith. That was probably the most awkward introduction we've had yet.

Eric: Yeah. We're getting worse in our introductions. Maybe slightly delirious today.

Alex: Maybe, where am I? What is the date?

Eric: And Alex we have a guest who's going to be talking about delirium today, right?

Alex: Yes we do have a guest, we have a guest on the Skype line, we have Tim Girard who's an Associate Professor at Critical Care Medicine at the University of Pittsburgh. Welcome to the GeriPal podcast Tim.

Tim: Thanks so much for inviting me.

Eric: And Tim, we're going to be talking about your New England Journal article about the use of antipsychotics in the management of delirium in the ICU, but before we do that we always start off with a song suggestion. Do you have a song for Alex?

Tim: I do, I'd like to hear Vice Versus by Switchfoot.

Eric: Great.

Alex: [Singing]

Eric: Nice.

Alex: Yeah I changed one lyric.

Eric: All right. Very nice.

Alex: That's why I was reading the lyrics beforehand.

Eric: Tim, why did you choose that song?

Tim: Well, you know when you guys asked me what song I'd like to hear I wanted to really be clever and come up with one that had to do with delirium but there just wasn't one that felt right about delirium and I just ended up choosing one that I thought was really beautiful I really enjoy this song. And also as I thought about it, I did think it made sense given that in the ICU we're often dealing with tragedy and in some cases with death and the song is really about how life is full of both the blessings and curses. And from the beginning to the end that it's all a complete package.

Eric: Great. That's a great song I'd never heard of that song.

Alex: I'd never heard that song before and I love it and so thank you for introducing us that song but, yeah, it does have some deep deep meaning as well as that nice drone in the drop D tuning for those of you out there who are guitar players and listening. Check it out it's fun, it's not too hard to play either.

Eric: I also think we have a bit of a vacuum cleaner in the background that's adding some nice reverb.

So Let's get started let's talk about this so you just published an article called “Haldol and Ziprasidone for Treatment of Delirium in Critical Illness” in the New England Journal of Medicine in 2018. Before we talk about the article, we always like to ask the authors like how did they get interested in this subject of delirium in ICU patients?

Tim: That's a great question I would say that I've been interested in delirium and in multiple complications of critical illness since I was resident, I was fascinated with the critically ill at the time and subsequently ended up going into that field. And delirium was a problem that was very common for critically ill patients but often was not given much attention in fact if you go back a couple of decades it was just labeled ICU psychosis by many and was considered sort the price of doing business in the ICU. And to give that line of thinking some credibility at the time, it was thought that the symptoms of delirium would just resolve and had no lasting effect. But a lot of studies have now shown that patients who are delirious during their critical illness even though the symptoms of delirium almost always resolve prior to discharge from the hospital, there are many adverse long term effects associated with delirium including long term changes in cognition.

So I became very interested in studying that, worked closely with Wes Ely whose one of the pioneers in this field of study for many years while I was at Vanderbilt and I've continued to focus on delirium and cognitive effects of critical illness after moving to the university of Pittsburgh a couple years ago.

Alex: So Tim you studied... You and this tremendous group across the country did this study comparing how Haloperidol and Ziprasidone and Placebo for treatment of delirium, and can you talk a bit about just the last sort of piece of background for my end is, are antipsychotics as a class prescribed, and Haloperidol in particular and Ziprasidone and why choosing Ziprasidone? Maybe starting with antipsychotics and Haloperidol.

Tim: Okay, yeah. There are several questions there, so you might have to remind me to cover them all. So first of all just with antipsychotics in general, antipsychotics for better or for worse are very frequently used to treat delirium whether agitated or not in the ICU. Now, why is that? I think there are a couple reasons, one is if you go back and look at the recommendations that were made in Clinical practice guidelines as early as the early 90s or certainly in 2002, there were guidelines published by the society of critical care medicine in 2002 in which Haloperidol was recommended as the first line treatment for delirium in a critically ill patient. Now, you could look at that and say well based on the dosing that they recommend, what they're really talking about is agitation but the bottom line is if you're looking for a guideline how do I manage delirium in the ICU that guideline which was by far in a way the most commonly sided guideline for delirium in the ICU recommended Haloperidol, based on no data, no data whatsoever in ICU patients.

And they were transparent about that, you know they weren't claiming that they knew that it worked, they just said look this is our best guess based on data primarily outside the ICU in the geriatric literature. And those guidelines I think had a major impact because when you look at observational studies there have been numerous observation studies that show that antipsychotics are frequently given to patients in the ICU.

Alex: Yeah, so I had a patient just last week in the ICU delirious and had been for I don't know, five days or so and met with the ICU team and they said, we finally started treating his delirium today with Haloperidol and I said wait a second, didn't the New England journal medicine paper just come out about that. It is incredibly common I think in our anecdotal experience and also it was what I was taught.

Tim: Right. And you know what's really interesting is there it seems to me that when I look at the data, you know and certainly I've got my own experience you've got your experience, we all know what we've seen but when you actually look at the data that have been published it's very consistent that you see about 40 percent in these observational studies, 40 percent of patients with delirium in the ICU whether hyper active or hypoactive or mixed picture received antipsychotics at some point during their ICU stay and then when you back up and say well what if we don't even ask the delirium patients let's just look at all comers to the ICU, there have been a number of studies that have looked at this and those studies have reported it's surprising how consistent the numbers are. Multiple studies not even considering whose delirious and whose not, have found that 11 percent, so one in 10 roughly of patients who just enter an ICU, one day or more in multiple studies receive an antipsychotic at some point.

Tim: But here's the part that I'm really concerned about, of those patients who receive an antipsychotic in the ICU, almost all of whom are naïve to antipsychotics. They don't come into the ICU on antipsychotics, and some of these studies have specifically excluded patients who are on it prior to their ICU admission. So you look at patients who had never received an antipsychotic, they come into the ICU, they have a one in 10 chance of getting an antipsychotic in the ICU and a quarter of those patients who receive it, end up being discharged on an antipsychotic. This has been reported in at least four different observational studies from multiple different parts of the world, and you look at that and there's no way you can attribute that to single dose treatments of agitated delirium because patients aren't being discharged from the hospital while agitated and delirious, we know that's the case and yet 25 percent of the patients who received an antipsychotic in the ICU go home on the drug.

So there's very clear evidence that we've had mission creep, what started out as a recommendation to do something for a patient whose got agitated delirium in the setting of an acute critical illness has gradually become a scenario where patients who have no longer have any symptoms of delirium, agitated or not, are being discharged from a hospital on antipsychotic. So one of the reasons we wanted to do this and you can talk about all the potential biological hypotheses that might explain why antipsychotics would work or why they might not work, and I those are important but even if there was no biological rationale we're looking at a practice that had become very, very common and yet was not supported by any placebo controlled trials.

There had been multiple small trials done that were all negative and our thought was look we need to know whether or not this has any benefit that we can detect and if it does then great, if it doesn't we need to do a trial that is large enough and convincing enough that will lead to a change in practice because if this practice is not beneficial for patients it needs to stop.

Eric: Yeah, and you know I get the Haloperidol, Ziprasidone I've never prescribed ziprasidone, I know some people do. How did you pick that? We asked the question of Twitter, what question should we ask you, that was one of the questions that we actually got.

Tim: And I think it's a really important question and it relates to... The answer to that question is just a little peek into what it takes to do a randomized control trial of a drug in the United States. So we went to the FDA as you need to do when you're going to give a drug off label in the setting of a randomized trial or via different route and in this case we were doing both, right? Because these drugs are not approved for the treatment of delirium so it's clearly an off label use and they're also not approved as an IV medication for any indication. So it was both a different route and a different indication so we needed to get approval from the FDA, so we submitted an IND and another thing that we really wanted to do in addition to giving it as an intravenous medication, because we think that that is the most common way to give antipsychotics at least initially in the ICU, as we wanted to do this as a double blind trial.

And to do it as a double blind trial giving intravenous drugs you would need to give an intravenous atypical antipsychotic and the FDA had no reservations, you know I guess I shouldn't speak for the FDA but my perception was that they had few reservations about us giving haloperidol because it's frequently done even though it's off label, been lots of papers published about it but they had some serious concerns about us giving an typical antipsychotic because there's very little data published on the safety of typical antipsychotics. Now I think that's changed in the time since we originally designed this trial, Olanzapine is now a more often given as an intravenous drug then it had been at that time but at the time they said look for us to consider approving a study that includes the use of intravenous aypical antipsychotic, you need to present us what existing data are available to tell us what the safety might be in humans.

And we pulled everything that we could find and the most of the data were based on Ziprasidone, which is somewhat surprising because like you said it's not a drug that we would frequently see used in this setting and yet that's where we're able to find safety data in humans for intravenous route of delivery.

Eric: And I learned something Ziprasidone not Ziprasidone, is that right?

Tim: You know what, I honestly don't even know which one either I don't know..

Alex: You say ziprasidone I say ziprasidone.

Tim: Yeah sounds like neither one of us should say... We just shouldn't even start using that word in the ICU at this point but...

Alex: So let's jump into the methods and could you just gives us like the 10,000th foot overview of what you did here?

Tim: Sure, so in general this was a randomized double-blind placebo control trial I think most listeners will be familiar with what that design means. One thing that was a little different about this trial design than many other trials, is that we attempted to, and we didn't succeed in every case, but we attempted to identify these patients as early as possible. In some cases we tried to identify in every case but in some cases we succeeded at identifying these patients before they became delirious and yet the randomization needed to occur when treatment was going to start. And treatment was going to start when the patient became delirious, so we actually had what was called a pre-randomization phase of our research activities.

Now, that phase wasn't actually part of the clinical trial because the clinical trial starts when randomization occurs and patients are randomly assigned to one of the three treatment groups but we had this pre-randomization period during which we were screening patients looking for patients who were at high risk for delirium and we used some pretty simple criteria, if you were mechanically ventilated in the ICU, or were in shock on vasopressors in the ICU then we considered you high risk. Of course, we had some exclusion criteria we didn't want to include patients who were at high risk for adverse effects, we didn't want to include patients who had server neurologic injury at baseline and there're some other exclusion criteria that are outlined in the paper. But we identified these patients sometimes even before they became delirious, they could be non-delirious because they were comatose and there was no way to assess for whether they were delirious or they could be non-delirious because they were alert and had normal cognition.

And our goal was to identify those patients to obtain informed consent from the patients or their surrogate decision makers as early as possible. For many patients that was at a time when they were already delirious and if that was the case we would just immediately randomize them into one of the three treatment groups and start delivering the drug in a double blind fashion. But for many of the other patients they were either comatose or had normal cognition at the time that informed consent was obtained but that didn't mean the trial started then, instead we would observe them for up to five days as long as they were still in the ICU and if at any time during that observation period they did become delirious that's when the trial would start for them, they would be randomized and then they would start receiving the treatment, the double blind study drug.

Eric: And so to be included for the study you had to have delirium based on the CAM, right? You were using the cam to...

Alex: CAM ICU.

Eric: CAM ICU and it didn't matter if you were hypo or hyperactive delirium either way you were enrolled in this study?

Tim: That's right. And there were two reasons for that, one... Two reasons for us including both hyper and hypoactive delirium. One was that, and I've already mentioned this data, there were a lot of data that showed that patients are getting the drug in both types of delirious states, hypoactive and hyperactive. And then the other is that the largest group of patients in terms of motoric subtype in the ICU are those that have a mixed subtype, so they fluctuate back and forth.

Eric: They don't read the manual that you have to either be hyper or hypo.

Tim: They don't. That's right, it does not read the DSM so-

Eric: And like what we said before even though it's hypoactive patients, they're still getting it because they're still potentially trying to pull out their lines and their tubes and there... Yet again not reading the manual of what they're supposed to be doing.

Alex: Well I think there's some push back here one of my mentees Pulmonary Critical Care fellow here at UCSF and he was saying that the UCSF critical care group feels like well what if they'd done the study in only patients with agitated hyperactive delirium might have been effective because those are the ones who come to our clinical attention the most and who we end up treating the most as opposed to the hypoactive delirium often sneaks under the radar it doesn't get treated.

Eric: Well, it's really the mixed that we see the most probably actually. Like think about all the patients that we see with delirium they're not just hyperactive all the time, they go from hypoactive usually usually to start off with some hyperactive. And like we saw the outcomes are still as bad.

Tim: In this conversation about the hypoactive versus hyperactive has really highlighted for me something I would have done different if I were able to re-write this paper, you know non of us can look back at a paper, I assume non of us I guess I can't speak for everyone else I will say from my own experience I don't have a single paper I've written that I would look back and say well that was perfect I really got every single thing right. And for this paper I feel like people have really focused on the breakdown of hyperactive and hypoactive in table one, but what I failed to do as an author, and I guess I can let all my other co-authors share some burden for this or some blame, but what we failed to do is to indicate that it wasn't just those 11 percent who had ever had hyperactive that was the number who had hyperactive at the time that they were randomized to one of the three treatment groups.

If you actually look over the course of the full study, it was more than a third, it was pretty close to 40 percent who developed hyperactive delirium at some point during the trial. And that is probably actually even an under estimation of the percent who had delirium at some point and the reason I say that is that those percentages are based on our twice daily assessments as a research team. So if, for example, a patient at 11 P.M. when we weren't there at the bedside doing our research assessment became agitated and the nurse who was taking care of that patient said well, this is a concerning situation we better do something about it, let's give a little more Propofol if that was the sedative their patient was receiving at the time. Then by the time we saw the patient again for one of our twice daily assessments in the morning as a research team we would not have found that the patient was delirious and hyperactive.

Eric: So what you need to do is actually do all the other previous trials in antipsychotics did which was don't even mention hypoactive or hyperactive delirium. Because like I actually went back to some old studies like that HIV study of like hospitalized HIV patients who got Haldol, the enrollment criteria was like did they have delirium? It wasn't did they have hyperactive delirium, then I looked at some atypical studies, like nobody enrolled patients with just hyperactive delirium in very quick review of literature, which is fascinating that all of a sudden we have this article, I won't jump to results, but potentially it shows that it doesn't work and then people say, no because they're not hyperactive but we don't even have literature to say that it works in those folks.

Alex: I think we should jump to results. Let's jump to results and we were talking about table one here which is who's in the study do you think that the people in this study are representative of people in ICUs around the country who are A, delirious and B, delirious and being treated?

Tim: Well I think that they are representative of the sickest patients who are delirious in the ICU, remember we only included patients who we considered high risk because they were mechanically ventilated or in shock. And if you look at the percentages by far and away, the most common reason that you're eligible in this study was because you were mechanically ventilated but there are a lot of non mechanically ventilated delirious patients in ICUs across the world, that are not really represented in this trial. So I guess if you wanted to make an argument that the trial doesn't apply to the non-mechanically ventilated patients who are not in shock, I think that that argument stands. I would respond by saying well, where are the data that do suggest there's a benefit in that population because I'm not aware of the data and I think at this point given what we do have in terms of published trials comparing antipsychotics with placebo in the ICU setting.

The burden I believe is now on those who are using these drugs to say well here's the reason, here's the evidence that supports this practice, and it should no longer be the default to give these drugs to delirium, agitated or not, when multiple trials, however imperfect they maybe, have yielded no evidence that the drugs help to treat delirium in critical illness.

Eric: And just to remind me so, your results, like big picture results that you found there was no difference in number of days alive without delirium or coma, right? There was no difference in multiple other outcomes including days of delirium, days with coma, really anything else we looked at right?

Tim: I want to be really careful with my words because it's important to be precise as a researcher, the way I would answer that question is we found no evidence of a difference. Now does that prove that there's absolutely no difference? No it doesn't. Remember what we did was we designed a trial, that was powered to detect a two day difference in number of days alive without delirium or coma, and that's a mouthful but really what the primary outcome came down to is how many days did you have normal cognition during the treatment period right? We treated patients for up to 14 days as long they were in the ICU during that time, and we measured as our primary outcome the number of days that they had normal cognition, that is they did not have delirium, they were not comatose.

And we powered the trial to detect a two day difference. Now, you could say, well then maybe there was a one day difference and you couldn't detect that and I think that's technically true. I mean statistically our data cannot prove that there was not a 0.5 day difference in number of days alive without delirium or coma. That said, if you look at the graphs, it shows you what the estimates are and it certainly isn't compelling in terms of suggesting maybe there was a difference but we just weren't powered to detect it.

Alex: No, they're not.

Tim: I can't say with any sort of statistical certainty or a statistical test to back me up that there is no difference, all I can say is we found no evidence of a difference.

Eric: Did you look at like... Another question from the Twitter verse and did you look at harmful behavior, like did it change like people pulling out lines, or tubes, or?

Tim: Yeah, so I think that's a really good question I would say yes and no, in terms of did we look at it. No, we didn't look at it directly in the sense that those events are very rare and we didn't directly observe the patients for 24 hours a day to detect that, it would be difficult to detect it purely by chart review because sometimes when those events occur they don't end up being documented in the chart. Certainly a self extubation is almost always documented, but pulling out an IV is probably almost never documented, so it would be hard to detect those kinds of events. However, the reason I would say, yes that we did assess for evidence of those events is that we took very great care to document all the sedating medications that these patients received not just the drugs that were given as part of the trial but also any drug that was given by the treating ICU team that could sedate the patient.

We tracked all of that and then we did a very careful analysis to find out if we could detect any difference in the amount of sedation that patients received. And we also tracked open label antipsychotic use, so you know one of the common questions we received is well, why did you decide a trial that allowed patients to receive haloperidol, when that was one of the drugs being studied? Well it's not that our protocol allowed that, it's that we did what is considered the standard of care when you design clinical trials and that is to do an intention to treat analysis and so we very much wanted to avoid any open label use of antipsychotics and tried on a consistent daily basis. We would go to these ICU teams, we would remind them every day that this patient is in the trial, here's what that involves, they should not receive any open label antipsychotics. We even provided the clinicians caring for all the patients that we're in this trial with an evidence based rescue protocol.

Tim: Now that was something that they could follow based on their judgment of what was best for the patient but it wasn't required in our words, the trial only controlled whether they received one of these three drugs, Haloperidol, Ziprasidone or placebo, which was saline. It did not control whether they received Propofol, or benzodiazepines or others but we did have a recommended approach to the treatment of agitation and the reason we did that, well there were two reasons one is because we wanted to provide these patients with what we thought was the best available care and then the second was that we wanted to avoid use of open label antipsychotics. And even with all those efforts there were patients who received some open label antipsychotics. The good news is that it was a very small amount that roughly 20 percent received at least one dose but when you look at the overall dose received as open label verses as study drug it was the open label doses were minuscule compared to the amount.

Alex: Compared to the amount that they received as study drug. Now this is a question that comes from our geriatrics journal club, the average daily dose of haloperidol was 11, right?

Eric: You can go up to 20.

Alex: You can go up to 20 but 11 milligrams a day was the average daily dose and I must admit I think I was the only person in the room who said they had prescribed doses as high as that the dail,y and no one else in the room had. So I'm interested-

Eric: Bunch of geriatricians.

Alex: No go slow.

Tim: This is one of those situations where if you're trying to be all things to all people you're going to end up disappointing everyone, right? We knew based on our own experience as well as talking to many, many intensivists that there was a very good chance that we would do a randomized controlled trial of antipsychotics and hear from the intensivists at the end of the trial, well of course it didn't work, you didn't give nearly enough. At the same time we also knew that the geriatricians would look at us and say, well of course it didn't work you gave too much. Those doses are dangerous and so I guess you could argue well maybe you should do multiple multi million dollar, multi-year randomized control trials some studying low dose, some studying high dose, or you could potentially do a single trial that studied multiple different doses.

And I would actually argue that we did in one way do that, we did study multiple different doses in the sense that we didn't have different groups that received different doses but we had a protocol that started at a low dose for every patient and then the dose was increased if the patient remained delirious and decreased if the patient was no longer delirious. So our dosing protocol rather it titrated the dose according to what the patient's mental status was, and so you could argue and we started actually at let's see five milligrams. I actually have to look back and remind myself because we started on-

Alex: It depended on age, yeah.

Tim: Right we started a lower dose for the older patients than we did for the younger patients but we started at a low dose and then we increased that dose until they got to the maximum dose, unless their delirium resolved, if it resolved then we would go back down. You know, for those who say, well these doses seem really high you could argue that if a lower dose had been effective we would have seen evidence of that and would not have ended up giving the higher dose.

Eric: I also think this fascinating because I feel like if you use lower doses like folks... And we see this with other trials, oh they did use enough.

Alex: Right they didn't push it hard and high enough.

Eric: Or if you didn't see a difference folks would say, oh because you didn't use Olanzapine or Quetiapine. Do you think it would be different if you used different drugs too here? Because we don't really have any literature that those drugs work any better.

Tim: Right. So I should acknowledge that there was one placebo controlled trial led by John Devlin at Tufts that compared placebo with Quetiapine in criticalyl ill patients who had all already received haloperidol. So it's a bit different of a design they enrolled patients who were persistently delirious despite receiving haloperidol and then to that haloperidol they added either quetiapine or placebo in a double blind trial. There were 18 patients in each group and they actually did find in that trial evidence of a benefit patients who were treated with Quetiapine had earlier resolution of delirium than patients who were treated with placebo. Now why is it that we don't all give Quetiapine? I think the reason is that we have-

Eric: Because we're using it for sleep for everybody already.

Tim: Well maybe so, and in addition to that we all have had certainly in critical care I think this is probably true of every field of medicine we have a long track record of early small positive randomized control trials that end up not being replicated when larger trials are done. So I know John very well and I think he and I see eye to eye on this, we actually just talked about this again at the European Delirium Association meeting in the Netherlands last month. We both looked at those data and say, well that's really interesting, that certainly suggests a larger trial should be done but I don't think either one of us think that's strong evidence that we should all be giving quetiapine and why is it that randomized... I'll quickly as an aside give a little lesson in interpretation of clinical trials.

Why is it that small trial is less likely to be replicated? Well because the smaller the trial the less likely randomization is to end up resulting in evenly distributed groups and even if you look at table one in that trial, you'll see that patients in the placebo group, there were a higher percentage of patients came from nursing homes, they were a little older. So you have 18 patients in each group and they may not be the same.

Eric: And I'm also looking that you did not have to have just hyperactive delirium for that study too.

Tim: No you did not.

Alex: Well let's stick with the hyperactive and the... So I know that in this study at the time of randomization we're talking small numbers who had hyperactive delirium but as you just said, at some point during this study of those enrolled something like 40 percent had evidence of hyperactive delirium. Did you do a subgroup analysis in those patients?

Tim: So what we did to analyze the effect on hyperactive delirium is that we measured delirium every day in every patient and we measured whether that delirium was hyperactive or hypoactive. And as I said earlier that was at the time that our research staff were assessing these patients so it didn't necessarily detect all hyperactive delirium but using those data we were able to compare not only the overall duration of delirium in the two groups or as I said earlier the primary outcome was the number of days without delirium and coma. So we could compare that but we also were able to analyze whether the duration of hyperactive delirium and the duration of hypoactive delirium were different, so that's not exactly the same as doing a subgroup analysis where you sort of just section off all the patients who ever had hyperactive delirium and all the patients who ever had hypoactive delirium. Instead, we did it this way, because as we've mentioned several times patients can fluctuate, they can be hyperactive at one point, hypoactive at the other.

Tim: So if you did a subgroup analysis how would you know, which of the two groups to put the patient in if they had mixed delirium, I guess you could do a third group that was just mixed but instead what we did was we compared duration of hyperactive delirium between the three groups and we compared the duration of hypoactive delirium between the three groups. And in both of those analyses we found no evidence of any difference.

Eric: All right. So my two last questions, the first question is does this change your practice and if so like how do you manage hyperactive or how do you manage ICU delirium?

Alex: Yeah great question.

Tim: Yeah so those are two different questions. The answer of the first one-

Eric: Sorry I've got three questions for you then!

Tim: Does this change my practice? I don't think that it does, because I had... well in 2010 we published the pilot trial that we ended up using to obtain the funding to do this larger trial and in that pilot trial admittedly a small trial, it's really only designed to show that it was feasible to do a double blind placebo control trial of this condition with these drugs but in that pilot trial there was really no evidence of a difference. And at that time I certainly was not convinced that there was no difference but I just felt well unless there's a strong reason to believe that these drugs work I think I'll focus on another approach …

Eric: So, your practice didn't change because you weren't doing it. I'm going to save my two questions because I'm going to bring back my first question is do you convince others that their current practice should change based on this study then? How do you convince other ICU doctors that... Because Alex is looking like I'm confusing him, how do you convince other ICU doctors that they should change their practice or should we based on this study?

Tim: I think that it should change their practice, and my major concern is that patients who are started on a drug even with good intentions in the ICU end up at risk for being discharged on that drug, that now you can easily say, well that's a problem with medication reconciliation, and a problem with transitions of care and that's all true, but the bottom line is that we in the ICU we need to think about the ways that what we're doing affect patients not only in that moment at that time that we're writing the prescription to start an antipsychotic, but we also need to recognize that it may have effects down the line. And 25 percent seems like an awfully high percentage to me of patients who I started on antipsychotic in the ICU with good intentions but now are going home on drug that has a black box warning from the FDA when used as an outpatient medication. So I feel like if I had strong evidence or even medium quality evidence that it was beneficial in the ICU that would be enough reason to start it.

Alex: Okay. I'm going to use up my last question here and this is a question we should always ask, what would they do in Canada? And my understanding from our journal club is in Canada, they feel like hey, we're not going to be able to treat this delirium but this patient is agitated and pulling out whatever, we're going to sedate them and that's what happens when you get to high doses of antipsychotics, right? You're really using them more for the sedative effect so rather than giving antipsychotic they give a sedative and they give a mild one like Trazodone or Trazodone?

Eric: Dang those A’s!

Tim: Well I guess that gets back to the question of how do I manage delirium and I don't think there's one easy answer because every patient with delirium is a little different than the last. But for one I try to do consistently across the board, the ABCDEF bundle and we probably don't have time at the end of the podcast to go through every one of the components in that bundle but it's very easy to look it up if you Google ABCDF bundle, you will find multiple websites with information about that. That is an approach to care that has components that have consistently been shown to improve outcomes for patients who are critically ill and when packaged together and delivered consistently a number of studies have suggested that it reduces delirium. Now does it necessarily treat agitation? Not necessarily but it does look like it improves cognition, reduces delirium overall.

How would I treat agitation in the setting of critical illness? Typically, I would sedate the patient, now that's assuming that this was agitation that I thought was truly dangerous to the patient. I will say in my own personal experience that I have come across many many patients over the years that one person on the ICU team may think he's having very dangourous agitation, but when I assess the patient I think okay, you know it's not ideal that they're agitated but they're not actually pulling out lines or disrupting care. It's just uncomfortable to see them in this state. And I think we have to acknowledge that many times we're giving medications whether it's antipsychotics or other medications that will sedate the patient because it's just not pleasant to see them squirming in bed and I certainly don't want that for my patients. But you have to ask well is the alternative which is sedating them sometimes to the point of drug induced coma better and most of the data suggest that it's actually not better to use heavy sedation even if it's prompted by agitation.

So certainly if a patient is exhibiting agitation to a point that it's dangerous then yes, I would use sedation. If they were not mechanically ventilated then I would use a sedative that doesn't suppress respiratory drive and IV haloperidol would be one that I consider specifically because it may help with agitation and it would do so without suppressing respiratory drive. More often than that though, I would consider Dexmedetomidine.

Alex: Okay, Dexmedetomidine. I wonder though is there a possibility that an unattended consequence of the study is of an increase use of restraints, you know physical restraints rather than a chemical restraint?

Tim: Yeah I hope that's not the case, I'm definitely not a fan, I would recommend against using restraints unless absolutely necessary. My experience is that there are a lot of ICU and hospitals where culturally the norm is to use restraints and to sort of justify it as an alternative to chemical restraints but when you actually look under the hood what you see is that they're doing both so that's problematic.

Eric: Well I want to thank you tremendously for actually joining on this podcast.

Alex: Thank you so much too.

Eric: I learned a ton.

Alex: And kudos to you, this is a tremendous amount of work we have to say. A tremendous amount of work kudos to you, kudos to the team based in Vanderbilt and the 16 centers around the country. Amazing amount of work.

Tim: Yeah thank you very much. And I really appreciate it being on the podcast, you guys do a great job. I find these podcasts really helpful and I know a lot of others do.

Eric: We're going to end a little with Alex singing but what's the song called again?

Alex: Vice Versus by Switchfoot. [Singing]

Eric: And with that I want to thank all of our listeners for joining us today. We look forward to having you with us next week for our next podcast and again big thank you Tim for joining us.

Alex: Thanks Tim.

Tim: Thank you I really enjoyed it.





Girard_Final (Completed 12/13/18)

Transcript by Rev.com Page 1 of 1