Subscribe
Share
Share
Embed
Communicable takes on hot topics in infectious diseases and clinical microbiology. Hosted by the editors of CMI Communications, the open-access journal of ESCMID, the European Society of Clinical Microbiology & Infectious Diseases.
[00:00:00]
Erin: Hello and welcome back to Communicable, the podcast brought to you by CMI Communications, ESCMID's open access journal covering infectious diseases and clinical microbiology. My name is Erin McCreery, and I'm the senior director of infectious diseases strategy at the University of Pittsburgh Medical Center in Pittsburgh, Pennsylvania, US.
I am joined today by four other fantastic CMI Communications editors to discuss the clinical trials presented at ESCMID Global 2026. This is part one of a two-part series because there was so much fantastic content. A lot of us at CMI Comms doing this podcast are trialists, are clinician researchers ourself, are practicing, in labs or in clinical care still.
And so we inherently are somewhat involved in a lot of these trials, whether we intend to be or not. And so as we go through these episodes, we will, if we comment on the data for the trial, whether we're the one presenting the trial or have [00:01:00] commentary, we will verbally state if we have any conflicts of interest.
Otherwise, all of our conflicts of interest will be listed in the show notes. Or if we have a genuine conflict, we simply will not comment on that trial. So with no further ado, let me introduce you to the wonderful editor team here. So first we have Josh Davis. You know him, you love him. He's an infectious diseases physician and the head of the Infection Research Program in the Hunter Medical Research Institute at the University of Newcastle in Australia.
He's also a clinical trialist. Josh, welcome to Communicable.
Josh: Thank you, Erin. Great to be here, and I'm very excited to be talking about late breaker clinical trials, one of my favorite things. and I'm now going to introduce Marc Bonten. Marc is a professor of molecular epidemiology at the University Medical Center Utrecht.
Did I say that properly, Marc? Very good. Utrecht. where he is currently vice dean of education. Marc's also a clinical trialist and a fellow editor on CMI Communications. Welcome, Marc.
Yes. Thank you, and [00:02:00] hi all. And I'm very, very pleased to be here again. I think it's the third or fourth year in a row that we're doing this session on the trials at ESCMID.
I am especially pleased to introduce Anne-Grete Martson, who is our newest CMI Comms editor, and she's making her Communicable debut today. Anne-Grete is an assistant professor at LECDR. No idea what it means, what it stands for, but it is at the Leiden University where she's leading the antiviral pharmacology research group.
Marc: She's also a clinical pharmacologist. Anne-Grete, welcome.
Anne: thank you, Marc, for the nice, welcome and, very happy to be here. I'm happy to introduce, Angela Huttner, who is an infectious disease physician in Geneva, Switzerland, and of course, our editor-in-chief at CMI Communications
Angela: Thank you so much, Anne, for the lovely introduction. I'm so excited to be here as well. I'll dive right in 'cause we have so much to cover. We will start with our usual get to know you question. So for this [00:03:00] week, if you were organizing ESCMID 2030, what city would you organize it in and why? Marc, you go first.
Marc: Yes, I've been thinking about this for the whole week now, and I made actually criteria. So it should, for me, be easily accessible by train with not too high risk for delays. There should be a reasonable chance of good weather.
There should be nice and affordable restaurants, and there should be something to spend the day outside of the conference. So that brings me to either Paris or Amsterdam. Pick your choice.
Angela: Ooh. Nice. But, I mean, I hate to say this, but it's, like, often in Amsterdam. Marc-
Marc: It's April ...
Angela: you guys always get it.
Marc: Yeah, well... '
Erin: Cause they have tulips. It's wonderful.
Marc: Yes, it's- If you're fancy tulip, that's- But Amsterdam actually- Never thought of the tulips for- Yeah ... for that reason, but- Are you
Erin: kidding? That was the first thing I did when it was in Amsterdam. I went up to the tulips. It was amazing.
Marc: People go
Angela: for flowers and, and- No, I fully agree.
Like, Amsterdam is practical and, yeah, easy to reach. [00:04:00] Very practical. Next up is Josh. What's your city
Josh: so I guess there's limitations now that ESCMID has become so gigantic. There were, like, 18,000 delegates this time. There's only a limited number of cities that can actually have the infrastructure to do it. So I don't know what cities can and can't, but I was going to say Paris also because I think Paris, for me, is a really good overlap between it's a big enough city to do it and it has the infrastructure, but it's also, you know, a beautiful place to visit, and it's romantic, and there's things for tourists to do, and I'm a Francophile, so I would say Paris.
Angela: okay, so you guys are very, very realistic people, but the entire premise of this question is totally counterfactual. So I did not limit myself to any realistic constraints. What I would do would be to have, and this is a total bias obviously I would have ESCMID 2030 in Italy in Trieste, which is the city on the [00:05:00] northeastern border of Italy.
and you guys have probably heard me I've talked about Trieste already once at least on this podcast. so it's not realistic because I don't believe there is a convention center big enough to hold so many delegates. but Trieste is an amazing city. It's not hard to reach.
The airport from Venice is, not far, and it's on the sea, and it has good cuisine, and it has beautiful places to go, and it's an underrated city, which I'm very biased about 'cause it is the city where my mother originally came from. so Trieste, Italy for the next ESCMID global. Can I, can I change my answer? If we're being impractical-
Yeah, we are ...
Josh: I, I would have it in Sydney because- Ooh ... I'm gunning for Australia to be considered part of Europe, right?
Angela: Like
Speaker 6: Canada.
Josh: Yeah. We're getting more and more involved in ESCMID. We love Europe.
We love ESCMID. So I want you all guys to come to Australia. Sydney's an awesome city, and we can hold big [00:06:00] conferences.
Angela: Love it.
Erin: Sydney makes a lot of sense from a logistics standpoint too. okay, so I also... I wrote this question, and then as you were asking it, Angela, I I was like, "Did I think of an answer?" But I did. so along the somewhat dream, 'cause I don't think structurally it could hold it, although I do think they have a 15,000-person conference center, so just shy of what's needed for ESCMID Global. but I would do it in Zagreb.
I adore Croatia. Ooh. And , I would actually do it in Split or a coastal city like Hvar. Nice. Like, let's just rent the whole island, you know? Ah. Let's just like take... Actually, I don't even know if there's infrastructure on Hvar for 19,000 people. But, I mean, that would be the dream.
Again, we're, we're dreaming here. Oh, my gosh. Yeah. To be on the sea, oh, my gosh, and in the olive oil vineyards and wineries. Ugh. I mean, what a dream. But practically, I, I don't know, the capital might play, and then it's not too far. And then if you want a vacation on the front end or the back end, so that's what I would do.
Angela: yeah. Nice. Definitely off the beaten track. I love it. Yeah, Zagreb is a beautiful city. I don't know Split. I've seen pictures. It's amazing. We can always dream.
Yeah, exactly.
Erin: Anne-Grete
last but not least.
Anne: Yeah. So, I'm totally [00:07:00] unoriginal, but I also thought of Paris. fits with, uh, Marc's criteria, I, I must admit.
But I actually thought Paris because, 2020 was canceled, and it's so sad. I feel like it was sad for them. They should have another chance.
Angela: Sad
Anne: for
Angela: us. I still
Anne: feel robbed. Yeah. But also them, I guess- Yeah ... that they spent a lot of time, organizing and everything. Mm-hmm.
Indeed. Was canceled quite last minute in the end, so
Erin: It's funny you say that- Good point ... 'cause every time I think of Paris, I think of how in March of 2020, we all were like, "Oh, this will only last for two weeks." And we kept-- I kept all my bookings for ESCMID Global. I was like, " It'll be fine.
Everyone's telling us this will be two weeks of lockdown, and then it'll be fine to travel." And oh, my gosh, how little we knew back then. we did end up canceling that trip, which was sad. So sounds like ESCMID Global organizers, you have a resounding vote for Paris, so. Okay. Let's get started with the trials.
So . In this first episode today, we are each going to review one trial, and then for part two, we'll each share one more.
So listeners, you're going to get a review of a total of 10 trials, which we [00:08:00] are very excited about. And Marc, my friend, why don't you start us off with the first trial?
Marc: Yes. Let's do that fast. So I will start discussing the fast antimicrobial susceptibility testing for negative bacteremia RCT, which was presented by Ritu Banerjee from Vanderbilt University in Nashville, and which actually was in a session organized by JAMA.
So JAMA organized a session on clinical trials presented by people from the United States in Europe, and they were publishing at the same time as the presentation was, each of the studies in their journal on their website. So that was really an event. Now, the FAST trial. Little background on how we practice.
We have a clinical suspicion of infection, we take a blood culture and usually start antibiotics, and then we wait for the results of the blood culture. If the bottle turns in positive, we make a gram stain, and then we have some guidance on how to treat the patient optimally, and then [00:09:00] we can start by adjusting treatment, de-escalation, escalation, waiting for the susceptibility testing results, which can take more hours.
Now, the hypothesis here was that if you reduce that time for waiting for susceptibility results, that that will improve patient outcome. So the intervention here was that there would be direct antibiotic susceptibility testing and identification with a new tool, the Vitek Reveal, and that, would be compared to the standard of care, which means subculturing the blood culture content, waiting for the bacteria to grow, and then do the susceptibility testing.
Now, the potential gains are obvious. you have faster information to give the appropriate antibiotics, which could improve patient outcome, and you could have information to faster de-escalate antibiotic pressure and thereby, let's say, reduce the selection pressure for antibiotic resistance.
The primary outcome of the trial [00:10:00] was the desirability of outcome ranking door at day thirty after the gram stain. And the door is a composite endpoint with, outcome parameters. I think m-many of you are familiar with that. And in addition to that, there were many, many secondary outcomes. The study was done in seven sites in Greece, India, Israel, and Spain.
And very interestingly, the study could be done as in six of these seven sites, there was a waiver for informed consent. So the randomization could be done in the lab, and the physician and patients were just not informed about that they were part of a trial. I've tried to do this trial in the Netherlands, but it would not be possible without an informed consent.
That's why we could never do this trial, and that's why I'm so happy that now we have the trial. In total, they randomized eight hundred and ninety-nine persons, four hundred and fifty to the intervention, fast, and four hundred and forty-nine to standard care. Now This was a typical hospital-based [00:11:00] population with a median age of seventy-three years.
Almost fifty percent of the patients came from Greece, thirty-five percent from Israel, and just over fifty percent of them had hospital-acquired bacteraemia, mostly coming from the urinary tract. And very importantly, this was a population in a high prevalence AMR setting. Thirty-s-seven percent of the isolates were resistant to cephalosporins, and nineteen percent were also resistant for carbapenem antibiotics.
So this is a, let's say, above average, certainly for Europe, resistant, setting, antibiotic resistant setting. Now, before anything happens, because of the intervention, people will be trial, patients will be, uh, will be treated with antibiotics, and the initial antibiotic treatment was considered effective in about sixty-two percent of all patients equally in both arms.
Now, what did the intervention achieve? Actually, there was a higher proportion of patients on what they considered effective antibiotic therapy at twenty-four hours. [00:12:00] It was eighty-four percent in the intervention group versus seventy-five percent, and that's for the total population. Now, for the cephalosporin-resistant subpopulation, the difference was larger, eighty-two percent versus sixty-three percent.
And for the carbapenem-resistant subpopulation, the difference was even larger. It was seventy-two percent versus forty-seven percent, but that was at twenty-four hours. And indeed, there were more stewardship recommendations made within the first three days, and more recommendations were accepted within these two days.
So there was definitely an effect of the intervention. However, that effect was already equaled out at day three, because at day three, the proportions of patients on effective therapy were equal in both groups and also for the subpopulation. So the window of benefit of this intervention lies between twelve and forty-eight hours after the Gram stain.
And that window of benefit is nicely illustrated in the survival curves up to [00:13:00] effective therapy. There, the lines are initially similar, and then they, let's say, they create some space, and then they come together, and then they create this window of benefit, which is larger if resistance increases.
Now, the question is: What does that window of benefit create for the patient? Well, and that's, in this trial, actually very disappointing for the patients because the door probability was forty-eight point eight percent, and actually fifty percent means that there is no difference in any, let's say, in the composite of the endpoints for patients, forty-eight point eight percent Going from forty-five point three to fifty-two point four actually also means the same, that there is no real outcome difference for patients.
And that's the primary outcome of the study. However, there were also twenty-seven secondary outcomes, and among these twenty-seven secondary outcomes, there was one in which there was a clinically significant difference that benefits the intervention, and that was the proportion of patients that were still hospitalized at day [00:14:00] thirty in gr-- subpopulation of carbapenem-resistant infections.
So in one out of twenty-seven, that was the positive signal. Now, to me, the study is very valid. It's a randomized trial, and I think it was done, at least presented and also described in the JAMA paper as executed, I think, according to the standard of care and the standard practice in which you need to do a randomized controlled trial.
It delivered a very, very precise effect estimate in the relevant patient population, thereby to me providing high quality evidence that FAST, as it was tested here, does not improve patient outcome, and that making decision based on FAST, on the other side, also doesn't hurt patients. It doesn't produce any harm.
Marc: what now to do with that conclusion? And, well, if you're in doubt, you could start by reading the editorial that was published at the same time. And actually, I, had to read it twice because initially I was not sure that [00:15:00] the editorial was on the same paper that I was discussing.
one of the lines at the end of the editorial says, " The FAST trial should move hospitals firmly in the direction of implementing rapid AST." And that left me completely puzzled because this implementation will come with a price, definitely, and there is no discernible benefit. So it will be very, very difficult to consider this to be a cost-effective approach.
Marc: And I was walking out of that session and walking towards the exhibition of the conference, where I see the, let's say, the multi-billion industry telling us that everything we must do is create fast susceptibility testing and rapid diagnostics. I-I'm a senior now in this field, but this is-- this has been so for thirty years now.
And this study Again, doesn't provide the evidence that we're really changing something for patients, and it could be, I think the study tells [00:16:00] us, again, that gaining several hours in changing antibiotics towards what we consider optimal treatment does not yield the obvious gains for patients that we are expecting.
That could mean that our empirical treatment is so good that changing that in a way that we think that makes it better doesn't make a difference, or the other way around, that it's not really the antibiotic treatment that, let's say, makes the patient outcome. That's more the underlying disease of the patient that determines whether the patient survives or not, how long that patient will stay in hospital.
So I think this is a very, very important study that raises serious questions whether all our, expensive approaches to, fasten our practices related on, on susceptibility testing is really worthwhile. And I'm very, anxious to hear what you think of this.
Josh: Well, for me, this is a really good example of why we need randomized controlled trials in [00:17:00] general, because it seems obvious before doing the trial that if you have a technology that, gives you the antimicrobial susceptibility results much faster and appropriate antibiotics faster, that would translate to benefit.
And in fact, the old kind of way of changing practice in infectious diseases, you know, 10, 20 years ago, people would've just done, studies based on does it give you the results faster? Yes, therefore, we should do it. But now the bar is higher, right? we're more and more experienced at doing trials.
Now we want to see things actually change outcomes, and this hasn't changed outcomes, which was, you know, it's quite disappointing, but it's a good example for me of why we need randomized controlled trials.
Angela: I love that your interpretation is so subversive. I think-
Marc: No, no.
Angela: Yeah. That's fair.
It's
Marc: following the data.
Angela: Well. That's
Marc: something else, that's something else than subversive.
Angela: You're completely- The editorially
Josh: subversive
Angela: [00:18:00] Yeah, yeah, yeah. I have to say I fully agree with you, and I would maybe even question you touched on this, Marc.
obviously empiric therapy is going to be good, right? That's why we choose it. Yep. So maybe the outcomes, maybe the targets here are not the right ones, right? Maybe they don't have to ask that question, "Is this good for the individual patient?" Maybe there could be such a thing as a randomized controlled trial that actually says, "Okay, is this good for the collective?
Is this good in the aggregate?" You know, because we do care about proper antibiotic use, right? I think they set themselves an impossibly high bar. Mm. How are you gonna show that,
Marc: but the antibiotic resistance signal is interesting because the large- let's say that the more extensive the resistance is let's say the larger the signal of an effect, which could lead to the interpretation- Oh, no.
We just have- Yeah ... we just have to wait until there is more resistance, and then this will improve outcome. But we're not waiting. We're changing our empirical therapy for that. Mm. [00:19:00] So if the resistance goes up- Mm ... the empirical therapy will change as well.
Angela: will be weakened, yeah.
Marc: So either we wait until we have no more possible- Yeah
changes in empirical therapy. But then of course we probably also won't have much options to adjust.
This trial will definitely not end the discussion and studies on how we diagnose and treat patients, but at least we have a new answer here. And with that, I would give, let's say, the floor to Angela for the next trial.
Angela: Yes.
I will be pivoting from diagnostics to treatment. I'm gonna tell you about the Cefmec trial. And first off, I've got to thank Emily MacDonald for the following notes on this trial. Full disclosure, she was meant to be co-hosting today, not me, but she ended up being on a flight right now somewhere over the Atlantic, so that's why I get to be here.
She did not leave us hanging, though. You may recognize her in some of these notes in the description of the Cefmec trial. So the full title of this trial is Randomized Clinical Trial of Meropenem versus Cefmetazole for Definitive Treatment of [00:20:00] Bloodstream Infections due to ESBL-Producing E. coli. First author on this trial is K.
Hayakawa, and the senior author is Yohei Doi. Another quick disclosure, this trial was presented only by poster, so our details are quite limited, and there, of course, was not this sort of presentation followed by discussion dialogue around its results. so here we go. The rationale for this trial, like the Peter Pen trial, whose interim results were presented this year at the trial run session and were also discussed in depth in our most recent episode, was to test a strategy to spare carbapenems, in this case, for the treatment of ESBL E.
coli. And as we know, the MERINO trial concluded that definitive treatment with pip-tazo, piperacillin-tazobactam, was not non-inferior to meropenem for essentially ESBL bacteremia. The interim results of the Peter Pen trial do not have the same conclusion. Again, go back and listen to our most recent episode.
But either way, we need alternate [00:21:00] options to carbapenems. So enter the antibiotic cefmetazole, . Which has demonstrated in vitro activity against ESBL E. coli. And in observational studies, it has also shown outcomes similar to those with carbapenems.
But of course, observational data are not enough. So I'm gonna tell you a little bit about this antibiotic because possibly you may never have heard of it. It's only marketed in Japan, where it was developed, in China, South Korea, and Taiwan. It was actually approved in the US a long time ago, but it was withdrawn from that market in the 1990s, not for any safety signals but simply for commercial reasons at that time. So cefmetazole is a semisynthetic broad-spectrum cephamycin.
It's often classified as a second-generation cephalosporin, and it's used parenterally only. It inhibits bacterial cell wall synthesis by targeting penicillin-binding proteins, -- in these four countries, it's used frequently for intra-abdominal infections, pelvic inflammatory disease, respiratory [00:22:00] infections, and soft tissue infections.
So you can guess that it has very broad-spectrum activity, gram-positives, gram-negatives, anaerobes, including Bacteroides fragilis. On the safety profile side, it can cause hypoprothrombinaemia, so It can prolong the prothrombin time, and therewith, it can predispose to bleeding.
It's also reported to cause delirium, and finally, it can cause a disulfiram-like reaction, so that means no cocktails while you are having your ESBL bacteremia treated. it's a lot like other beta-lactams. There are dose adjustments to be done for reduced renal function. So onto the trial now that you've had your cefmedazole primer.
CEFMECH was a non-inferiority trial comparing cefmedazole to meropenem. This was an entirely investigator-initiated trial funded by Japan's Society for the Promotion of Science, which is essentially the equivalent of an NIH body. So the study took place in twenty-seven hospitals in Japan, and it was open-label. Like most of these [00:23:00] trials, patients were already on antibiotics at the time of enrollment, which could occur up to seventy-two hours after the index culture. Randomization was one-to-one, and it was stratified by infection source.
No further details available. Again, we're going off a poster and by hospital site. Both drugs were prescribed as one gram IV every eight hours, with each dose infused over, quote, "thirty minutes or more." And the duration was determined by the prescriber but was between five and fourteen days. In this trial, the pre-specified non-inferiority margin was ten percent for a primary outcome of mortality, and a sample size of a hundred and ninety-six was needed for eighty percent power.
So in the end, two hundred and seven patients were randomized, but we don't actually know out of how many screened. That's not marked. There were different populations, There would be an ITT population, but we don't have those results. We are given the modified ITT-1 population, which had one hundred [00:24:00] and ninety-nine patients out of those two hundred and seven randomized.
Among these one hundred and ninety-nine, a hundred and two were randomized to cefmetazole and ninety-seven to meropenem. There was also a per-protocol population defined and an MITT-2, uh, but today we'll just focus on the results they presented, whi-which were that MITT-1 population. So these patients were older, median ages of eighty-one and eighty-three in the cefmetazole and mero group respectively, and around forty-five percent were male.
About sixty percent of the infections were community acquired. A low percentage came from the ICU, around ten percent, and the PIT score was low at one. So many infections, just like in Peter Pen and Merino and Astarte, were urinary in origin. It's pretty classic for these gram-negative, Drugs. From the MITT-1 population, two out of 102 patients died in the cefmetazole arm, and five of 97 patients died in the meropenem arm.[00:25:00]
So this translates to a risk difference of -3.2% for mortality, very clearly making the 10% non-inferiority margin. There were several subgroup analyses, but the number of deaths were so few, it seems less relevant to get into them, as many of them even had zero counts for the outcome. the investigators did look at retrospectively at a 5% non-inferiority margin, and the p-value for that was .005.
on the safety side, there were very few serious adverse events, fewer than 10% per arm, and none of them were deemed due to the study drugs. So to conclude, this trial was quite small and had few events. Patients were older and comorbid but certainly not critically ill. The investigators did successfully demonstrate non-inferiority of cefmetazole to meropenem with a 10% non-inferiority margin and likely even with that 5% non-inferiority margin.
So the trial has a lot of strengths. [00:26:00] Obviously, it's beautifully randomized. It's nice multi-center, 27 sites, and it uses a hard objective outcome of mortality. It of course has some weaknesses, the biggest one being its open label design. But I can tell you placebo controlled trials are really hard to do on an investigator-initiated budget.
And here the investigators rightly chose essentially the most objective primary outcome possible. I would say right now the other weaknesses are likely mostly related to reporting, which is really not the investigator's fault as they were restricted to a poster. So I'm really looking forward to reading the full report.
In the meantime, I would say it seems like we now have another study suggesting that definitive therapy for ESBL bacteremias may not always require a carbapenem, which is exciting for stewardship, of course. So while some of these alternative antibiotics that are being tested may not yet be available in all countries, . Results from randomized trials like these can be used to [00:27:00] advocate to bodies for access, which is why we are giving this trial some airtime here. A little more access in other parts of the world would be good for everyone. Guys, what do you think?
Josh: I've got a comment that's not really about the trial, although I think it was a great trial.
Very interesting. I don't know if we're supposed to be saying this, a bit spicy. My children would say a bit of tea about this, but, um- Why was this given a poster? we had a- Mm-hmm ... we had a major trial run session where we talk about ESBL, gram -negative trials, both of which were either half finished or post hoc.
This was finished and it showed something. so Mm ... I, I kind of feel like apologizing to the investigators, but not that it was anything to do with me, but that this got given a poster. So now that I've got that off my chest- Yeah ... let's keep
Angela: talking
Josh: about the
Angela: trial. Oh my gosh. I do agree with you, Josh. The ESMO Global Program Committee, which we none of us are on, clearly doesn't agree with us.
but I think it's important to make this point my best guess, and again, I can't speak [00:28:00] for them, I'm not on the program committee, my best guess is that they're just like, "Look, this is an antibiotic that's available in a very restricted part of the world.
there has to be relevance, clinical relevance." and I would assume that- Yeah, I push- That's where the thinking we think But so is
Josh: temocillin I
Erin: was gonna say- Yeah. Temocillin. So I'd push back on that- Yeah
a lot because the antibiotic that was given the platform is only available in two countries or three countries currently, I think- Mm-hmm ... which are very small. You're talking
Angela: about temocillin.
Erin: Yeah. It's only available- Yeah ... what, in Belgium and no offense to Belgium, but like I think Japan's bigger. Yeah.
So, I think that it is a shame, a crying shame that this was a poster considering there was a session on ESBL bacteremia treatment. So kudos to the authors for this, and I think that would have loved to unpack this more if we knew more than just what's on the poster, so.
Ouch. . Moving on. Josh, you're up next.
Josh: Yeah, thanks. so the trial I'm going to present is called the COBRA Trial.
That was also presented in another late-breaker session, which was focusing on surgical site infections. And, uh, [00:29:00] quite a lot of people I've spoken to have identified this one as the trial that they think will change their practice. so this was looking at patients with acute cholangitis, and it was randomizing them to one day of antibiotics, that being the intervention, versus the current standard of between four and seven days of antibiotics after drainage, of their biliary tract, um, in obstructed cholangitis.
So there are guidelines that exist, for example, the International Tokyo Guidelines, they recommend four to seven days. I've checked the Australian guidelines are similar after a successful ERCP drainage of cholangitis. There was a randomized trial published in twenty twenty-four comparing four with eight days, showing that four days was not inferior, and there's observational data suggesting that even shorter than four days is just as good.
so that was why they did this trial. This was done in thirty-one sites across the Netherlands. It was an open-label, government-funded [00:30:00] investigator-initiated trial. the patients had acute cholangitis. They had to have had, ERCP guided drainage of their biliary tract, which was generally obstructed by a stone, and they had to have had their fever resolved.
they excluded patients with immunosuppression, neutropenia, staph aureus bloodstream infection, Pseudomonas bloodstream infection, a liver abscess. and of note, they did not exclude people with bacteremia in general. So think about it, they're getting one day of antibiotics. Some of these patients had bacteremia, about thirty percent of them actually.
They were randomized on one-to-one basis, and it was stratified on the causative organism, gram-positive, gram-negative. there was a blinded outcome assessment committee, and the primary endpoint was clinical cure, and that was defined as being symptom-free by day fourteen. And having no relapse or death by day 30.
So it was a composite of those things. there were a number of other [00:31:00] outcomes, including 90-day relapse and others. it was a non-inferiority trial with a margin set at 7.5% absolute risk difference. they, based the sample size on a 90% expected cure rate in both groups, and that gave them a target sample size of four hundred and sixteen.
they end up having four hundred and ten patients randomized out of seven hundred and forty-five screened, two hundred and five in each group. the mean duration of antibiotics actually received was what they were randomized to get. So it was one day in the one-day group and five days in the longer group.
and the bottom line was that there was no difference between these two groups in the primary outcome. So it was 95.1% versus 93.7% in the short versus the long group. so that met the non-inferiority, definition. The absolute risk difference was 1.5%, but it was approximately zero, basically.
there were a number of subgroup [00:32:00] analyses, and , in the subgroups, the point estimates all favored the shorter course including those with bacteremia, gram-negative bacteremia. and in terms of the secondary outcomes, again, they all favored the short course or were no different, and that included 90-day mortality.
The only thing that was different was antibiotic-related adverse effects were more in the long group, so 16.6% had at least one antibiotic-related adverse effect in the long group versus 8.3% in the short group, and that was, statistically significant on the frequentist design that they used here. So for me, this was a really nicely done trial.
it very clearly showed that one day of antibiotics was non-inferior to the current standard, and that it's kind of audacious to do that in patients with gram-negative bacteremia, but you know, it worked. to me, the key issue is around generalizability because, a lot of patients I see with acute cholangitis don't get [00:33:00] an ERCP done.
Uh, and if they do, they don't get it done quickly. So I didn't mention, but, on average, the patients had their ERCP about 48 hours after admission to hospital. and often in my hospital system, they'll wait around for ages to get the ERCP or just not have one at all. and since I've seen this trial, I've been rounding on patients the last couple of weeks, and nearly none of the cholangitis patients I've seen would have been eligible.
So the last three I've seen, for example, had underlying sclerosing cholangitis, and they just get kind of recurrent, cholangitis without an obstruction. so I think if I do, you know, have the typical patient with a cholangitis and a stone that have an ERCP, it's unblocked and they're quickly getting better, I would stop their antibiotics just like we do for appendicitis as soon as their appendix is out.
another caveat is that this is not published yet, so we only know what was presented, but it has been submitted to a journal, and the protocol has been [00:34:00] published. But so I look forward to seeing the full paper, but I think from what I've seen of this, this should change practice.
Erin: Yeah. I think this is incredible, for two reasons.
One, so I would say after if patients are not bacteremic, as soon as their ERCP is performed, we stop antibiotics. So that tracks with our practice in that regard. We actually don't even give the day after unless it's like their ceftriaxone dose fell at 3:00 AM or something and when they came out of procedure.
but if they're bacteremic, we give seven days from source control. so this is, I think, immediately practice-changing and wild to save patients up to six days. And I think, Angela, we've said this before on other podcasts. I think it goes to what we've been saying in that there's a lot of debate about oral step-down therapy for uncomplicated gram-negative bacteremia.
Can you use a beta-lactam? Do you have to use a fluoroquinolone? And I actually think none of it matters. I actually think all of the oral step-down is irrelevant exposure and that you really only need-- I've been saying three days of therapy, like IV, three days of IV therapy for gram-negative bacteremia, but now maybe we only need one day from source control-
which is [00:35:00] insane, but also very cool. So kudos to the authors.
Angela: Yeah, I fully agree. I think this is exciting, and
we shouldn't extrapolate too much. But come on. I mean, this is hugely meaningful for abdominal infections where you have source control, where you know there are no remaining abscesses or, you know, really this starts to change the discussion. This could be very paradigm changing, I think.
Erin: And we
Angela: know people
Erin: from the Netherlands are awesome because we have two people,
Angela: two people,
Erin: two of the five of us. So kudos
Angela: to you. They're very bold. We love it.
Josh: Kudos to you. They're, they're very good at using not much antibiotics, people from the Netherlands. We need to learn from them. Yeah.
Angela: , I would make one minor note on the actual design trial, which, you know, I think it's cool that they got away with this, but I wonder if reviewers will give them a hard time for choosing non-inferiority margin of 7.5%, when that primary outcome includes mortality, right?
and in the end, you have a very small event rate. So you're basically-- With that non-inferiority margin of 7.5%, you're basically saying, "Okay, we will, you [00:36:00] know-- If our event rate is 3%, we'll be fine with a mortality that ends up being double," right? always a game, right? You can't have a really narrow margin and actually afford the trial you wanna do.
But, I wonder if they're gonna have some trouble with that, honestly. Yeah.
Josh: I do agree with that, Angela. But on the other hand, it's pretty standard, even in mortality-based non-inferiority infection trials to use 10% and even 20% some trials as non-inferiority margins. It's quite unusual to use what clinicians-- as clinicians we think it should be, like 5% or less.
Yeah. Because you end up getting a giant sample size then.
Angela: But, hmm, nowadays, Josh, I think really even trialists don't honestly accept 20% margins. That's just- No, not 20 ... you know, that's just insane. But,
Josh: but 10 is pretty s-- the standard-
Angela: Still ... most commonly used. Yeah. I, I-- Yeah. Yeah. I see both sides, But anyway, we need to move on.
Josh: I'm gonna hand over to Erin now, and she's gonna tell us about, the DOTS-PK trial.
Erin: all right. DOTS-PK, what a joy. I'm excited to [00:37:00] present this trial. I do wanna verbally say AbbVie, the company that owns dalbavancin, at least in the US market, I think globally as well, they have a database called the INFORM database. That's a Gram-negative resistance surveillance database.
And in 2025, I gave a talk at the Making a Difference in Infectious Diseases conference on the INFORM database, and AbbVie did pay my travel for that talk. So, just to be upfront about my relationship with this company. I have an emotional relationship with dalbavancin in that I adore it, and so I will be will be frank in that this is something we've really shifted practice to in my institutions because I think it's really wonderful for patients to not have to have a central line and get an antibiotic every single day.
I think just bottom line up front, if you ask someone, "Hey, do you want this thing in your body and have to inject yourself all the time?" Most of us would say no to that question. And so oral antibiotics, long-acting antibiotics or things that help patients avoid that traditional standard care, I think is where we should be thinking, in terms of the patient experience.
And actually, the SNAP trial [00:38:00] team in New Zealand, has like a really beautiful commentary on oral antibiotics in CID published last year by Walls et al. that has patient quotes about how they feel about taking oral antibiotics versus receiving IV antibiotics. If you guys haven't read that, it's a beautiful paper that really talks about the patient experience.
So with no further ado, the DOTS pharmacokinetic data was presented by Tom Holland from Duke on behalf of the ARLG in the United States. The DOTS trial, for those who may not be familiar with it, was published in JAMA in September of 2025. This looked at dalbavancin versus standard care for complicated staph aureus bacteremia, and they found-- They enrolled 200 patients in that trial, randomized one-to-one, and they found a similar risk-benefit ratio with receiving dalbavancin or standard care.
they said dalbavancin was non-inferior to standard IV therapy by clinical efficacy. I think it was 72 versus 73%. They did have, Josh, that 20% non-inferiority margin though, and that was some very reasonable criticism of that trial is that [00:39:00] the non-inferiority margin was massive. You could like sail a ship through it.
Probably too big, 'cause we-- that is a significant clinical difference. Uh, but they had, you know, similar efficacy in terms of percentages, similar safety. It notably was not-- Their primary-- They tried to power this for superiority, and it was not superior per a door analysis. So that's the background of the DOTS trial.
But The cool thing is that they had a pre-planned secondary analysis of pharmacokinetics in patients that received dalbavancin, and their objectives were to characterize the pharmacokinetics of total and unbound dalbavancin in patients that had complicated staph aureus bacteremia.
And this is the first study of actually infected patients. We have lab data before this and model data, hollow fiber models. But as we always say, people are not hollow fiber infection models, so people matter. Your immune system matters. Your sickness d- impacts how you clear drug. So what did they do? They took plasma PK samples of total and unbound drug before the first dose, 10 minutes after the first dose, [00:40:00] at six, 12, and 24 hours, at day eight before the second dose, and then day 22, 44, 70.
They used a nonlinear mixed effects model, and they applied that to simultaneously characterize total and unbound dalbavancin in terms of concentration time-dependent data. They wanted to identify patient factors associated with variability in dalbavancin exposures. Again, everyone's getting the same dose.
some patients got, a little bit of a dose adjustment if they had renal dysfunction, but most people are getting 1,500 times two, and they wanted to evaluate, here's the real important one, whether dalbavancin exposure was associated with clinical success. So I think this is awesome. we have animal and hollow fiber infection model data showing us that exposures out to 21 days after a dose of dalbavancin may be above a minimum inhibitory concentration of eight.
And so that was reassuring that lab data, because practically when you're trying to administer this to patients, not every patient can come back at exactly day eight for an infusion appointment. And so those, like [00:41:00] operationally and, you know, in my health system, those lab data gave us some relief that if a patient said, "I can't come Monday, can I come Wednesday?"
We were like, "Yes, you're not gonna die between Monday and Wednesday. You're gonna have enough dalbavancin." There's some wiggle room in this like day zero, day eight dosing regimen. But now we have people data, which is even, better, right? And, so again, before the DOT trial, all the PK was based around this n- neutropenic murine thigh infection model, which was a six-day model.
And it really showed us that with these two doses, you should have therapeutic concentrations for 42 days of total and free drug, and that assumed a 93% protein binding. So what did they find in this PK analysis? So of the 100 patients in the trial that were randomized to dalbavancin, 97 of them had samples available, and then 93 patients were event-free through the first few weeks and evaluable at day 70.
So this is the PK sampling from 93 patients. And they found that one, we were wrong about the protein binding assumption. [00:42:00] Dalbavancin is actually highly protein bound, greater than 99%. Over 90% of the samples had a fraction unbound of less than 1%. So this is significantly protein bound. Indeed, later unbound concentrations in the patients they sampled were undetectable.
At day 42, 73% of the patients had a level below the limit of quantification for unbound concentration, and at day 70, 98% had a level below the limit of quantification for unbound concentration. So super protein bound. Creatinine clearance had really no association with the fraction unbound, but albumin, as expected, had a modest inverse relationship with fraction unbound.
So the lower your albumin, the more drug is free, the more drug is cleared, the more likely you are to potentially have underexposure, which is the story we see with cefazolin and ertapenem. This is increasingly becoming incredibly important in the pharmacotherapy space in assessing patient's albumin and then TDM accordingly.
So here's the real important part to end with. The [00:43:00] relationship between dalbavancin exposure and clinical success. Does it matter? I think this is the Holy Grail question for any antimicrobial. Is there a MIC exposure relationship? Is there exposure outcome relationship? And so among patients in the DOTS trial who experienced clinical success, dalbavancin concentrations were higher on average than patients who experienced failure in both their AUC exposure and their day 22 single point concentration.
They saw a similar trend with unbound concentrations, and they found that efficacy increases as dalbavancin exposure increases when they divided patients into kind of these thirds. And so the highest level, like the highest chunk third, had the highest rates of efficacy, and then there was a middle group and then a lower group.
And to take it one step further, they found that there was a cut point they could declare. So 32 micrograms per mil seemed to be this cutoff. They had 30 patients that had a drug concentration greater than 32 at day 22. Of those, [00:44:00] 29 out of 30 experienced clinical success. Contrast to the 60 patients who had a drug concentration less than 32 at day 22.
Erin: Of those, they had a 25% absolute lower success rate. Only 43 patients or 68% experienced clinical success. So What do we do with this data, these data? I think one, we know that dalbavancin is 99% protein bound, and albumin's going to matter when you're treating these patients. Higher exposures are equal to higher success, and they really didn't see any safety signals with the higher exposure, although the numbers are quite small when you're looking at 30 versus 60 patients.
You're not gonna get an AKI signal, but no notable glaring safety signal with a higher exposure. The parent DOTS trial again gave two doses, because that was the data we had from in vitro models, and then there was also a small RCT in osteomyelitis that used 1,500 milligrams times two doses as kind of definitive therapy.
So that's where the dosing [00:45:00] regimen came from. but clinical failures in DOTS were relapses that occurred later. So the direct quote from the presenting author here was that it was biologically plausible that it is due to lower exposure at the end of therapy, these failures that they saw.
So the real question is we give two doses now, can we do better? And dalbavancin versus, you know, standard vanco or daptho for six weeks was similar. Can we do better? Is there better than a 25% failure rate in staph aureus bacteremia, which is kind of what we see consistently. so to me, honestly, these data were immediately practice-changing insomuch as I have two pharmacists that work for me that run a transitions of care program.
And when patients are unhoused or have no insurance or significant barriers to receiving antibiotics outside of the hospital, we often are leveraging this, agent to try to minimize the amount of appointments and resources they need. And, we've just been doing two doses, and I literally texted my team from [00:46:00] ESCMID Global and said, "We need to start checking day 22 concentrations," which are fortunately available in the United States now.
Those weren't available in the past couple of years. And so for countries that don't have this assay, I mean, I think you do need to pair this with TDM. Don't just blindly be given doses. We have a pretty clear cut point from these data. But I said we should start recommending day 22 concentrations, and then based on that concentration, potentially recommending more doses patient, specific infection specific.
So I think these are really cool data, and interested in y'all's thoughts.
Anne: Yeah, I think it's super cool they measured the unbound because obviously we don't see that that often. I'm just curious, like what's the trend in unbound concentrations? I haven't read the paper but if it's, linear between the unbound and bound, and if it's nonlinear, then you would expect, that the unbound is more critical to measure.
in clinical practice, I don't see anyone measuring unbound. I mean, that's gonna be really hard to do. But, yeah, I would be curious to [00:47:00] see if it's like a saturable situation or if it's, a linear relationship between them.
Josh: I knew you'd be excited by this one, Erin. It's like a, perfect one for PK people to nerd out about, right?
but it's also cool that it's shown correlations between, drug levels and clinical outcomes, which is really interesting. One of the questions this left me with, which I've kind of brought up a few times over the last few years about dalbavancin is, what would they have found if they had used oritavancin in the trial rather than dalbavancin, which has a significantly longer half-life?
that's what I've got no conflict of interest here except emotional. I like oritavancin. that's what I use in our clinical practice here. We generally only give one dose, rather than two doses or possibly three doses now might be needed. So, that's an unanswerable question because there's very few data, but it would be interesting to know.
Erin: It is a super good point because when we talk about the long-acting lipo- lipoglycopeptides, [00:48:00] you have oritavancin on the table. You know, some people have the stewardship argument that oritavancin has activity against VRE and dalbavancin does not. So in a way, some of us like save eravancin for VRE it's like how we used to save- But do you
Josh: actually use it for that?
save
Erin: vancomycin. Yeah. Yeah, we will. So we have-- I mean, you don't-- you have less persistent VRE, but we've used oritavancin for particularly liver transplant patients and, once or twice we've done that Q48 hour for three doses lead in, of the MD Anderson. I think there's like two case reports in total on this PK assessment.
I don't know how accurate it is. but, think it just comes down whatever drug you have on formulary. But it's actually, it's kind of like the cefditofibo/ceftaroline argument. Like now we have an RCT for one, no RCT for the other. Are they interchangeable? No. Like we know, we know that drugs in the same class are not interchangeable.
Yeah. But we want to interchange them. You wanna take whatever your formulary agent is ' which honestly probably just comes down to whatever the best contracted price for your institution is, and then say like they're the same, but they're not the same drug. So it's, it's [00:49:00] important. We want, you know, ideally RCTs for all of the antibiotics.
All right. Anne, you wanna bring us home with the last trial of this first episode?
Anne: Yeah. I'm really excited to present this, maribavir, real world data. So this was presented at the late-breaking research from the Lancet, new interventions for more established pathogens, presented by Annalisa Pavlikniati from Barcelona, Spain.
So this trial, was called Maribavir for clinically significant cytomegalovirus infection in hematopoietic cell transplantation, a real world retrospective international study of the infectious disease working party of EBMT. So maribavir was approved by EMA in Europe in November 2022 for, treatment of RR or relapsed refractory CMV infection or disease after solid organ or, hematopoietic stem cell transplantation.
It has shown limited toxicity, especially compared to the other drugs in the market, and app- [00:50:00] approximately ten percent resistance, has been shown. so the study aimed to report clinical experience with in HC- HCT, patients, and it's basically a retrospective registry study in children and adults, who received maribavir post, stem cell transplantation after EMA approval.
So, they included patients with a minimum of 12 weeks of follow-up and maribavir indication, treatment of clinically significant CM infection. And if used before authority approval, These were excluded. The primary endpoint was CMV resolution after 12 weeks, after maribavir start, and secondary was the reason, for discontinuation, overall survival, and no relapse mortality at 12 weeks, since initiation.
and response to therapy was basically defined, as maribavir just discontinued with no need for additional anti-CMV treatment. And treatment failure was [00:51:00] then defined as, viremia progression requiring therapy change, addition, viral rebound, et cetera. So they recruited 118 patients, in 37 centers, but they do talk about courses.
So from these patients, it was 126 courses, which majority were male and, had acute leukemia. only nine of them were children, so mostly adults Uh, interestingly, 53% of patients received letermovir prophylaxis, which is kind of new practice in the last few years for this, population. And, the donor negative recipient positive CMV was 39% of the patients, so that would be our highest risk group.
the donor positive recipient positive was 50% of patients. Um, and the indication for maribavir was then either preemptive therapy, for 19% of the population or second-line preemptive therapy 48%. [00:52:00] reason to use, maribavir was refractory or resistant CMV for 51%, 25% because of neutropenia.
renal impairment for 12%, and other toxicity to previous antiviral a-agents was 10%. And as expected, the previous antiviral treatment was either valganciclovir for 40% or foscarnet 18%. now we're gonna look at some of the results. So, either the response on, preemptive therapy or during CMV disease.
So response for the, preemptive therapy courses in 109, courses was 81% was g- response. So that was a very, very nice outcome. Failure was only in 17%, 2% lost to, follow-up, and basically no response failure was only 1%. but CMV disease, During, preemptive therapy was in 5%, so five patients, five courses.
response for CMV disease, that was in 17 courses. There was 70% [00:53:00] response, 12% of these, 17, courses had progression on, maribavir, and 18% were lost to follow-up or not evaluable. resistance was actually in the no response courses, the 20 courses was tested in only six courses, and from those, only two had the resistance in UL97.
the reason for continuation was then 76%, was end of planned therapy. And, overall survival at 12 weeks from maribavir start was 82%, and no relapse mortality was 15%. So basically, maribavir showed to be effective treatment option for clinically significant CMV, in, stem cell transplantation, high response rates with limited toxicity.
so the authors and presenters, suggested maribavir use, beyond the EMA-approved indication should be further explored. I really like to see this study because obviously it's real world experience with maribavir. We have used it very little actually.
actually we always [00:54:00] use it, when our backs are against the wall, so when the patients are already really poorly and then we have lost those patients actually. So I've heard some practices that people start with foscarnet and then when the viral load is coming down, then they continue with maribavir.
Anne: So that's all like anecdotal experience, still. But, I think this is a nice, initial, overview of maribavir use. it is retrospective and of course not a huge amount of patients, but, I think it shows pretty promising, results actually. I don't know if anyone else has some ideas on this.
Anne, is this different to what was seen in the registrational trials of maribavir, the kind of
Anne: response
Josh: rates
Anne: they're seeing? Yeah, no. The registration trial was of refractory and resistant CMV as well, so it's the same indication. So they don't actually explore. they just kind of said, "Yeah, we should explore also maybe as a first-line treatment."
I guess that's what they mean. Yeah, I would be hesitant to do that actually because I know ganciclovir is really effective.
Erin: Yeah, I mean, I have another emotional [00:55:00] conflict of interest because I love CMV. It's my favorite- The same.
Anne: It's
Erin: my favorite infection. it's so great. It's so cool. the drugs are so neat, but the drugs are also terrible.
And so I think, you know, maraviroc coming to market was such a huge moment for patients. I think, you know, an oral effective therapy for a terrible infection was just so tremendous. So seeing it expand in patient populations and use is really exceptional. We save it too. I mean, maraviroc has a very low barrier to resistance.
You only need one genetic mutation to get resistance, and so it's not super great for patients with high viral loads and, and whatnot. So ganciclovir is still treatment of choice. So then if patients can't tolerate Gan or Valgan, we switch pretty readily, and then it's obviously become relatively first line for ganciclovir-resistant CMV.
Anne: Yeah, and that's usually very- Pretty expensive. Yeah, it's expensive. Is it? Exactly. Yeah. Yeah. It's ex- expensive here as well.
Erin: You know what else is expensive though? Giving seven electrolytes a day and having monitoring for IV foscarnet and ending up on ECMO and [00:56:00] dialysis or whatever the heck we do to patients because we destroy their bodies by trying to give them foscarnet.
So I think the cost-effectiveness of maraviroc is probably more than said itself.
Josh: Touché.
Anne: Yeah, the question is should we start it earlier? Because now- Mm-hmm ... I feel like we give them like in the last resort. Here we see also very low resistance, so that's a good thing, I think. I also see that they have a good response, so I'm just curious the ID with high viral loads, what would be your outcome there?
Angela: sounds like it's a perfect segue for a prospective randomized trial, right? I mean, we always have our little rule. Retrospective studies should not change practice, but they should be the sentinel, right?
They should be a big glaring warning sign saying-- Or, in a good way warning, hopefully, saying, "Wait, wait, wait, we see something." Yeah. We need to go further with our, trial. [00:57:00] People
Anne: are so hungry for this data. I saw, like, people just presenting case series with Maraviroc in the conference.
because there's just not enough real world data. So I'm really happy that- Mm-hmm ... this is out there now, we just don't have enough experience.
Anne: all of them are on-- most of them are on PET, so preemptive therapy.
That means that those viral loads are very low.
Erin: Yes.
Anne: Those viral loads are very, very low. They're like around 1,000-- hundred-- in the hundreds even-
Erin: Correct ...
Anne: sometimes.
Erin: Correct. Yeah. And, we don't use Maraviroc for high viral loads, but
Anne: anyway. No. Exactly. So this whole [00:58:00] population is going to-- like largely is not going to have a very high viral load, I expect.
Okay. So I think that brings us to the end of the first episode, right? These are the trials we were going to cover in, this one. So I'll close off by saying thank you so much for the conversation, and thank you, for listening to Communicable, the CMI Comms podcast. As mentioned, this was part one of a two-part episode, and we'll be back with more trials soon.
Josh: This episode was edited by Katie Hostettler-Oi, and the theme music was composed and conducted by Joseph McNade. Now, for the first time, this episode was not peer-reviewed. At our editors meeting recently in Munich, after looking at all our past episodes and how they did, we discussed whether to continue with peer review or not.
Given that there had been no findings of content errors or unstated conflicts of interest to date as a result of peer review, and also the worry we're creating unnecessary work for our peers, we decided [00:59:00] to go forward without peer review. But if you feel strongly about this, please let us know, and you can write to Angela by email.
this episode will be citable with a written summary referenced by a DOI in the next eight weeks, and any published literature we've discussed today can be found in the show notes. You can subscribe to Communicable on Spotify, Apple, or wherever you get your podcasts, or you can find it on ESCMID's website for the CMI Comms Journal.
Thanks for listening and for helping CMI Comms and ESCMID move the conversation in ID and clinical microbiology further along.