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.
Communicable E67- Q. catchup
===
[00:00:00]
Emily: 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. I'm Emily MacDonald, general internist and infectious disease trialist coming to you from McGill University in Montreal, Canada, and associate editor at CMI Communications.
Today's episode belongs to our recurring series called The Quarterly Catch-Up. As a, a reminder, every three to four months, we have CMI Comms editors pick one to two articles each that have been published in the past three months. And we'll give you our take on the studies and catch you up because we know everyone has very busy schedules, and it's hard to keep track of all the science being published.
we really aim to keep these episodes short. So as I always say, hop on the treadmill, get ready for your commute. We've got you covered with a rapid-fire update on the latest research. And unlike some of our other summary [00:01:00] episodes, like the annual Top 10 or our post-ESCMID episodes, we might not always focus on clinical trials, so these episodes are also a little bit about what the editors find interesting.
You can look forward to some variety in the science that we'll be presenting in these quarterly catch-ups. So today I'm really happy to welcome my co-host, Angela Huttner.
Angela: Hi, Em. I'm really excited to be here because you and I, we have never hosted an episode together.
Emily: I know. Incredible.
Angela: I know, in two years or more, so I've been really looking forward to this. I would note, interestingly, like Quarterly Catch-Up is supposed to be for all science and not just trials, but I think by virtue of the two co-hosts today , we did end up with trials.
Emily: Okay. So, so today happens to be all trials.
I'm gonna start us off. I have two studies to present, and I have a bit of a common theme in mind in that these two trials are looking at what's an acceptable treatment strategy when we factor in efficacy, [00:02:00] harm, safety, convenience, and cost, okay? So I'll be talking about OPTION and about POET too. So, I'll start us off with OPTION.
It was published in New England Journal of Medicine Evidence this summer. Johnson et al are the authors, and it addresses a very common problem. So what's the best treatment antibiotic regimen for recurrent C. difficile infection? And this is interesting because guidelines are increasingly favoring fidaxomicin for recurrent CDI, but the certainty of evidence is actually pretty low.
And another strategy that we use all the time, which is a prolonged vancomycin taper and pulse, has actually surprisingly little randomized evidence behind it. So OPTION was a randomized double-blind three-arm trial conducted through the US Veterans Affairs system and enrolled three hundred and eight patients, and this is important, with either first or second recurrence of CDI.
So these were not incident cases. These people had all recurred. Patients were randomized to one of three strategies. Okay, so the first [00:03:00] strategy was just standard vancomycin, a hundred and twenty-five milligrams four times a day for ten days. The second strategy was fidaxomicin, two hundred milligrams twice a day for ten days.
Or the third strategy, which was a vancomycin taper and pulse regimen, which started with the same ten days of vanco and was followed by a three-week, taper and pulse treatment Okay. The primary outcome was sustained clinical response at day 59. So resolution of the initial diarrhea without recurrent diarrhea through day 59.
I'm emphasizing this because We're not talking about microbiologically proven recurrence. This is slightly different, right? This is you had your C. difficile, you got better, and then you didn't have recurrent diarrhea. But of course, we know that a lot of people with C. difficile can get diarrhea for lots of other reasons, often older adults or immune-suppressed people.
So we just- we keep that in mind for the primary outcome.
Angela: And we are happy that they're randomized.
Emily: Yes, exactly. The, results, so the [00:04:00] main findings are sustained clinical response in 58.6% of the vancomycin taper and pulse group, 58.6, compared to 44.1 with standard vancomycin. So big difference there.
Absolute difference of 15 percentage points. Here's the kicker. Fidaxomicin was also 44.1%, so exactly the same as 10 days of standard vancomycin. So on the surface, pretty compelling win for vancomycin taper and pulse, but there is a little statistical hiccup or issue that arises. They made two primary comparisons, and very responsibly, they pre-specified a more stringent significant P value of less than .025.
And the P value for the taper and pulse versus standard vancomycin was .04. So the confidence interval excluded zero, but the trial technically did not meet its pre-specified criteria [00:05:00] for superiority. Yes. Secondary outcome tells a similar story. So they looked at freedom from recurrent CDI, and that was about 65% with taper and pulse and 49% with standard vanco and 52% with fidaxomicin.
So again, like, it looks quite good for taper and pulse. But the authors appropriately conclude that they didn't formally establish superiority statistically with either of the regimens. So- clinically, it's very hard to ignore the signal. vancomycin taper is cheap, it's familiar, it produced a 15 to 16% absolute improvement in sustained outcomes, whereas the standard course, vedoxo, performed essentially identically to standard vanco.
So if you just look at option alone and you don't look at, for example, other evidence that is out there, it doesn't maybe definitively tell us that the taper and pulse is best, but it definitely raises some very interesting questions about what we should be recommending and perhaps paying for in recurrent CDI.[00:06:00]
My first question for us to discuss and for our audience to think about is, is this really a negative trial? First of all, it's an inexpensive treatment. The number needed to treat is seven, and the 95% confidence interval excludes no effect. But the multiplicity adjusted P value, threshold of .025 was not met, so is that enough to not change your practice, to change your practice?
Was this really a negative trial?
Angela: All right. Can we go upstream of that? Because I'm a bit stuck on the fact that the time point for the primary outcome was day 59, and this is not from end of treatment, you know, in each arm. This is day 59 from randomization for everybody, right? So is this not giving a huge advantage to the taper arm because those people were on treatment for three times as long as the other two arms, so they just had less time to fail, right?
I mean, we all do take it as a given that vancomycin does work when you're on it, and the question I think that a [00:07:00] lot of patients would have, and their physicians, is not just how do you respond to it while you're on it, but how long are you good afterwards, right? are you protected from recurrence and for how long?
Emily: absolutely. Is there a certain amount of immortal-ish time for the vanco Mason pulse and taper group? You can recur on pulse and taper. We did see it, and maybe we'll talk a little bit about taper V, which was the trial that, my group ran for first episode or first recurrence, and that looked at a vanco pulse and taper of four weeks versus 14 days of vanco.
And people in the vanco and pulse and taper group could recur, but it was extremely rare. So there probably is an immortal-ish time period there for the longer treatment.
Angela: So immortal time bias.
Emily: Yeah, I mean, you couldn't achieve the outcome or it, it would be hard to achieve the outcome of recurrence in the pulse and taper group because you're on treatment for longer.
Angela: So those patients in that taper group are essentially protected. They're not being exposed to this potential outcome as much as the other patients [00:08:00] in the other arms.
Emily: Absolutely.
Angela: That's really something I stumble on.
Emily: Yeah, because, you know, your alternative also then if you said, "Well, I'm gonna follow these people from the end of treatment for the same period of time," then you would also...
disadvantage one of the groups as well. So the alternative of saying, "I'll follow for 56 days after the end of treatment," when we know that over time your risk of recurrence goes up, means that you would shift over by two weeks the risk of recurrence for the group that's on a longer treatment.
So at the end of your 56 days in your fidaxo group, for example, in your pulse and taper, you'd have an additional 14 days that you could recur in your pulse and taper group. Your option, is either you disadvantage one group at the end or you advantage one group at the beginning.
Angela: I don't think it's that simple because I do think we all have understanding that as long as you're on it, there is an effect, right?
You are being protected. If you attach two, three more weeks to their follow-up, they have more [00:09:00] time to fail, then you're penalizing them. But you're right. I mean, if you truly have, like, incredible equipoise and you don't think there's any protective effect, then yeah, I guess you're penalizing them if you have everyone have equal follow-up times post-treatment cessation.
But anyway, to get back to your real question, I have to be honest, for fidaxomicin it doesn't look good.
Emily: That does lead into my second question is, which is what happened to fidaxomicin? Why was sustained response 44 in fidaxo and 44 in standard vanco? And freedom from recurrent CDI was not really very different either.
It was about 52% for fidaxo and 49 for standard vanco.
Angela: Mm-hmm. I mean, I think probably the makers of fidaxomicin would say, "Wait a minute, you know, you're giving us a very high rate of recurrence, but you did not do due diligence. You're using an outcome that is purely clinical. You know, you don't have microbiologic confirmation.
These people probably get diarrhea a lot. They might be, more likely to report the littlest thing."
Emily: Absolutely. I think you could say, like, the [00:10:00] primary outcome may be a disadvantaged fidaxo. One thing I wanted to bring up was that in the original fidaxomicin trials, they looked at recurrence, for 28 days after completion of therapy.
That is quite a short timeframe to follow patients, and the difference between fidaxo and vanco, you see it early in those trials, and it does start to wane even as early as the early 28-day follow-up. And then, yeah, we saw in, Taper V that the taper that we gave patients, the difference between the taper and the standard course, if you follow the Kaplan-Meier curves, they start to approach each other.
And so some of these treatments, , to get at your point, may be delaying recurrence as opposed to clinical cure prevention. I think those are my main points. I do think we're making scientific progress. I don't think we can say that fidaxo is out. I, I don't think we can say that at all.
But the durability of its advantage is, I [00:11:00] think, less certain than we sometimes imply. A lot of times we just say "Oh, fidaxo is better," but it's a very short follow-up period for the original trials, and now we start to see that maybe, the durability as you go out further is not as strong as we've previously implied it to be.
Angela: Yeah.
Emily: And yeah, now we're running a trial Taper V2, which is vanco taper versus fidaxo, but we're following out to day 120.
Angela: Nice.
Emily: We're gonna really see, like, are these people cured, or are we just delaying their recurrence?
Angela: isn't like the concluding paragraph of the option trial-
Emily: Yeah
Angela: isn't their concluding paragraph literally that we need bigger trials that essentially, you know, go further, and you're doing it.
Emily: Yeah, we have to follow them for longer. I will say, and I've discussed this a lot, maybe delaying is not such a bad thing.
Angela: Hmm. It's not. I think patients would be happy to delay,
Emily: yeah, like maybe you don't want your C. diff to recur while you also just got over pyelonephritis. You know, like the two hits, [00:12:00] you're older, you're frail, you're already dehydrated, maybe you have a bit of kidney injury. It doesn't help to get your C. diff like right after you've been admitted for your gram-negative bacteremia.
Like maybe it's better to push it out six to eight weeks if you are going to have to have it. Okay, so this leads me into the second study that I'm talking about today, which is POET 2. Very exciting, published in the New England Journal of Medicine in August. I chose it because it challenges something that we almost never question in endocarditis, which is why do we treat everybody for four to six weeks?
So the original POET trial famously showed that selected clinically stable patients with left-sided endocarditis could transition from IV to oral antibiotics. POET 2 asks a different and perhaps even more provocative question. Once a patient has responded appropriately, and we can talk about what appropriately means, do they actually need the rest of the antibiotic course at all?
So this was an international open label randomized trial of 508 adults [00:13:00] with left-sided infective endocarditis. Three pathogens, Staph aureus, Enterococcus faecalis, and Streptococcus grouped together. Importantly, these weren't patients at the beginning of treatment. So before randomization, everybody had to receive a pre-specified two to four weeks of appropriate antibiotics, and they had criteria based on the pathogen.
So for example, Enterococcus faecalis, you had to have at least four weeks of antibiotics. Uncomplicated staph, two weeks. Complicated staph, four weeks. So it was somewhat protocolized. so then at that point, if you met the criteria for stabilization, 255 patients were randomized to this response-tailored strategy, stopped antibiotics, and then 253 continued conventional therapy, to complete the usual four to six weeks, depending on the pathogen and whether or not it was complicated or not.
They had two primary outcomes. They had an efficacy outcome, which was [00:14:00] a, a win of a pick, which was antibiotic, free survival, so the number of days patients were alive and not receiving antibiotics for endocarditis or bacteremia during, the following six months. Not surprisingly, if you were randomized to less antibiotics, you had less antibiotics.
So yes, great choice of a primary efficacy outcome. The response-tailored group had substantially more antibiotic-free time. Yeah, so it was a median of 183 days versus 169 days. So if you got randomized to the strategy of a shorter duration, you had about 13 additional days alive and off antibiotics. But let's look at the safety outcome.
that was the other outcome. I think that's the really important one because yes, you could randomize people to less antibiotics and they can get less antibiotics, but do they die? Do they get complications? I think that's what we're really interested in clinically. So the safety outcome The safety outcome was a composite.
It was not a hierarchical composite, okay? [00:15:00] So each outcome in the composite is equal. It counts equally. And the, the composite that they had was death, unplanned cardiac surgery, or symptomatic embolic events within six months. And they tested this for non-inferiority, and the margin was 7.5%. So originally the safety outcome had death, unplanned cardiac surgery, or symptomatic embolic events within six months and relapse. but then relapse was, removed without looking at the data.
And yes, this is tested for non-inferiority with a 7.5% margin. Okay, this outcome occurred at 8.2% of the response-tailored group, so the, the group that got shorter duration, versus 10.7 in the standard duration therapy. So that's good. The absolute difference actually favors shorter treatment by 2.4 percentage points, and they meet the criteria for safety for non-inferiority.
Sounds like a fairly convincing shorter is better story, [00:16:00] but wait. When we look at relapse of bacteremia or endocarditis, it occurred in 5.1% of patients who stopped early versus 1.6 in the standard therapy, and that difference just in that group alone was actually statistically significant.
Angela: But what were the, the absolute numbers?
Because you get down to some very small numbers there.
Emily: Yeah, so in the relapse you get down to some very small numbers. So I don't have the exact numbers, but it's something like 13 versus 5. Yeah, it's small. But statistically significant, just to say. If you kind of boil that down, you save 13 antibiotic days per patient without increasing death, surgery, or embolism, but approximately one extra patient will relapse for every 29 patients who you stop early in.
That's kind of the balance that we have to decide, uh, do we find that acceptable and do patients find that acceptable?
Angela: Again, I have to stop upstream and say, wait a minute, and you're saying this is statistically significant, meaning the [00:17:00] chance that this was a chance finding is very small. I do get that.
If somebody had sufficient resources and tried to reproduce this trial, I have such a strong suspicion that they would not come up with the same findings. But I get it. I mean, we obviously have these rules. We have statistical significance because we, want to give ourselves a certain certainty or probability when we discuss results.
I get that.
Emily: Yeah. Well, okay, one, one point I wanna bring up is- 20 years ago let's say giving someone six weeks of antibiotics versus four meant a PICC line, it meant staying in the hospital, it meant IV antibiotics in this group But now, as POET 1 proved, that additional two weeks doesn't need to be ceftriaxone in the hospital.
That could be two weeks of amoxicillin at home. And so if you're a patient and you say, "Okay, I could triple your risk of relapse," sorry, I know [00:18:00] that's relative risk, but you know, let's say for every 30 patients that I do this to and I shorten the course, one of you is going to relapse, or you could take two weeks of amoxicillin additionally at home.
I think a lot of patients would say, "I'm doing fine on these pills, like I'll just take another two weeks of oral antibiotics to prevent that chance of relapse," which is not huge. it's small, but it's, it's measurable and maybe the intervention is not so harmful at the point where you've switched to pills.
I might accept a longer duration of treatment if the treatment isn't so burdensome.
the longer duration of treatment used to be very burdensome, but now if it's just a matter of me taking another two weeks of pills and I'm tolerating them to prevent relapse, I would probably as a patient accept that. But if it meant staying in the hospital two more weeks for IV therapy, I would probably not accept that.
I think there are carefully selected patients where we can shorten the duration. if there isn't an oral option, for example, and they've met [00:19:00] all these criteria, like the CRP is coming down, it's resolved, the procal is negative, the TEE looks okay, like there were a lot of stringent criteria.
Then yeah, I think in a carefully selected population, you could shorten the duration and take the risk of relapse. But if there's an easy, cheap option to take at home that has very few side effects, maybe you're not going to accept the risk of relapse in that population.
Angela: I, I hear you. I tend to agree with you, and like I said, I think because these results are there, they're documented very publicly for, for a long time to come, it will be very hard for people to shorten.
And then don't forget, and this is something you've written a lot about I think, the pressure of sort of local dogma and, you know, local culture and, you know, it, it's going to take a very special clinician, a very confident clinician who's going to say, "Well, actually POET 2 tells me I can shorten," right?
I think, but I do fully agree with you. I actually do agree that a, a real specialist has an opening in [00:20:00] there for certain patients to shorten the therapy. And I also agree with you on the other side. Longer therapy today isn't the, the prison sentence that it was-
Emily: I mean, It reduced mortality to get people out of the hospital and, onto pills. Yeah. And, this was not like a, randomized to shorter versus longer trial, right? Like, this was really trying to find carefully selected patients in whom you meet these criteria could you shorten. It's very different than seven versus 14 days or four versus six weeks.
It does make sense, I think, to move towards, you know, finding patients who are, and determining duration of treatment based on some definition of, of clinical stability and improvement as opposed to us wanting an easy thing to remember, which is endocarditis gets treated with six weeks.
Angela: This is our eternal dilemma, complexity versus give me an instruction and I'll do it.
Emily: I'll turn it over to you .
Angela: All right. Actually, all four studies we're gonna talk about in this episode are all [00:21:00] essentially duration studies. This is kind of the thing right now, right?
Like, I don't seem to be able to escape it. The things people can do with these studies can be very creative and very clever, and therefore I'm gonna talk about PEDICAP as the lead in because yeah, very, very simple concept looking to reduce, durations, but the way people can do it is so nice. So PEDICAP is a study that came out, quite recently in The Lancet, and it's entitled Oral step down optimal drug and total duration of antibiotic treatment in African children hospitalized with severe community acquired pneumonia. A factorial randomized control trial.
Emily: Gotta love those Lancet descriptive titles. Very precise.
Angela: So this trial was run by an international team led by Julia Bielicki and Mike Sharland out of St. George's University of London. Full disclosure, Julia works between St. George's and Children's Hospital of Basel here in Switzerland.
So she is a [00:22:00] good friend of mine. And though we have not actually gotten to work on that many projects together, I am a huge admirer of Julia. And Mike Charland, another disclosure, Mike worked closely with my husband at WHO as they were developing WHO's Aware Antibiotic book. Maybe more importantly, Mike is just really funny.
He just makes me laugh way too much, and if you can do that, it will bias me towards you anytime. Oh, and then one final disclosure about this trial before I actually begin talking about it. We actually presented this trial on our very first Communicable episode in May 2024. This is because its preliminary results were presented as a late breaker at Estimate Global back in 2024.
All we had then was an abstract, and it's really just an interesting trial, so I thought it was worthwhile discussing it two years later. Finally, we have the full paper and it should hugely change clinical practice. So as its title says, it was a randomized trial that was built to ask this [00:23:00] question: Do we really need to give five days of intravenous antibiotics to children with severe community-acquired pneumonia?
Because that is what WHO recommends in official guidelines.
Emily: I did not know that. I wasn't aware.
Angela: And, and what people also may not realize, depending on where they live in the world, WHO guidelines tend to be followed not by high income countries because they're doing their own guidelines, but they are closely followed by low middle income countries.
And those are just the countries where it's insanely complicated to give five days of IV antibiotics. So this was a recommendation that was really not evidence-based, certainly not based on randomized trials. it was very expert opinion driven. so it was like high on the list of targets for let's produce the evidence for this recommendation because this is costing people.
This is costing African families to keep their kid hospitalized for five days in a row. And this recommendation, by the way, still exists in the first iteration [00:24:00] of the Aware Antibiotic book. They upheld it very specifically, though pending further evidence. So they're waiting for the results of this trial as well.
PediCAP included children aged two months to six years hospitalized with severe community-acquired pneumonia. I'm just gonna call that CAP from now on. These kids had severe pneumonia, but they couldn't have complicating factors with it. The recruitment happened in 13 hospitals across five sub-Saharan African countries, and children had to weigh between three and 30 kilograms, and they had to have a point-of-care C-reactive protein over 10 milligrams per liter.
So it had to be at just above the norm, to get into the trial. the pneumonia was diagnosed if there was difficulty breathing plus hypoxemia or central cyanosis, or if there was simply severe respiratory distress, or if there was a general danger sign. and WHO has a list of danger signs, inability to drink or breastfeed, vomiting everything, convulsions, or lethargy to the point of [00:25:00] unconsciousness.
So the interventions were, what I find quite interesting, . Included children were randomly assigned in a five to five to one proportion to either A step down? Oh, wait, I can't say step down with UM.
Emily: Good catch. Yes, I know. This is literally in the article.
Angela: So anyway, these kids were randomized to either oral switch from that IV antibiotic that they were started on to oral amoxicillin alone. So that was one option. Or they were randomized to amoxiclav, either of these drugs, if you were in either arm, was given twice a day, and the comparison between those two intervention arms was a superiority comparison with the idea, maybe you should be adding the clavulanic acid, to cover bacteria that have beta-lactamases like Haemophilus.
And then finally, there was this third arm, and that was randomized five to five to one. That was the one arm, and that was control. That was the standard of care according to WHO. That was the five [00:26:00] days fixed duration intravenous only treatment, and that was a non-inferiority comparison.
The Oral amoxicillin arms are compared to each other in a superiority comparison, but they're each being compared to control which is IV amoxicillin alone as a non-inferiority comparison.
So they get randomized to these three groups, right? But then there's a second randomization. So the children who are in the antibiotic PO switch arms, whether it's amox alone or amox-clav, they then get randomized to different durations.
Because mind you, the control arm, IV antibiotics, that's fixed at five days because that's the WHO recommendation, So there's no further randomization there. But the two oral switch arms get further randomized to either a duration of four days, five days, six days, seven days, or eight days.
So what were those criteria for oral switch? It was basically clinical judgment,, when the child had [00:27:00] clinically improved according to the clinician and was able to take oral antibiotics, that's when they did the switch. So the primary outcome was the proportion of readmission or death at day twenty-eight, and to be very faithful to what they write, "Ignoring intercurrent events of change from randomized treatment and restricted to children with known day twenty-eight status."
If I turn that phrase over in my mind, I'm reading that to mean this was very much an ITT, an intention to treat analysis. And they are quite upfront about any missing data. They ended up not having very much missing data. They do the analysis on everyone who was randomized, no matter what happened after randomization, and of course, in those people in whom they have a day twenty-eight outcome, and they've simply let you know what the numbers are.
Children who didn't have an outcome, they're just not in the denominator, but they are m-listed very clearly as missing. So the non-inferiority margin against this intravenous-only arm was ten percent. Uh, there's a long list of [00:28:00] secondary outcomes, both effectiveness and safety, that people can read in the article.
So just briefly, some statistical considerations. For that superiority comparison between amox-clav and amox alone, a thousand children needed to be randomly assigned one-to-one to those two intervention arms to provide, I think it was eighty-one percent power to detect an eight percent absolute reduction in the primary outcome from what they assumed would be around fifteen percent.
So they were assuming about fifty percent readmission in the amoxicillin group, standard of care, essentially. Fifteen percent is what the readmission rate is with amox IV. Yeah, that's kind of the baseline. so they were hypothesizing that you could get down to a nine percent readmission rate, a nine percent event rate, if you added clavulanic acid.
they put in a little assumption of five percent loss to follow-up. They needed to include eleven hundred children to end up with five hundred and five hundred and [00:29:00] one hundred. and in those Or switch arms of 500 each. They were then, of course, gonna be randomized to five more duration arms.
So you're ending up with about 100 children in each antibiotic duration iteration.
So to the results. They screened 2,248 children for a final sample size of, 1,101. So they did a lovely job. Interestingly, out of those 1,147 children who did not get into the study, the reason for 1,008 of them was having a CRP below 10.
So these were kids who were really unlikely to have a bacterial pneumonia. But I think they did a nice job there, honestly. All in all, quite good external validity, and I think a pretty well-selected population. Who were these patients? The final included cohort, young, and genuinely acutely ill. So their median age was 13 months.
62% of them were on oxygen at [00:30:00] randomization, 90% had severe respiratory distress, and, about 12% had severe malnutrition. So straight to the primary outcome. All-cause readmission or death at 28 days was really low compared to what they expected, but this is kind of the trial effect as we all see, right?
You do a trial, and suddenly everyone behaves really well. Yes. So all cause readmission or death at 28 days was 6.9% in the amoxiclav arm and 5.6% in the amoxicillin arm. That makes an adjusted risk difference of 1.3% percentage points. So no superiority if you added clavulanic acid.
That's the message there. Yeah. So WHO was not wrong in being quite targeted. Yes. Great for the world over, right? We don't need to be adding clavulanic acid. I think that clavulanic acid makes a big difference in terms of microbiome disruption and, emergence of resistance down the line.
Emily: And maybe like GI side effects too that can [00:31:00] sometimes cause readmission, like, you know.
here in Montreal, we always think of it as it causes more C. difficile, and we have a big C. difficile problem in Quebec.
And So I am aware of of it being problematic for that reason, you know. So I, try to use amoxicillin without the clavulanic acid when I can. But I do get a lot of interesting looks, like when I'll use it for E. coli bacteremia, I'll use amoxicillin without the clavulanic acid when I have the sensitivities.
But people wanna give both of them.
Angela: Yeah. They feel that it's stronger. Yeah, yeah. You're right to do that with the E. coli because it's just a beta-lactamase, and if the E. coli doesn't have it, it doesn't have it. And clavulanic acid in and of itself doesn't really have that much antimicrobial activity, so it's only there to help the amoxicillin.
so that was a huge conclusion already.
Emily: Yeah, we can switch to amoxicillin. We don't need the clavulanic acid.
Angela: Yeah. You can feel b-good about giving amoxicillin alone for s-severe pneumonia requiring hospitalization.
Emily: Yeah.
Angela: All right. So next big conclusion. Both oral step-down strategies easily met the pre-specified 10 percentage [00:32:00] point non-inferiority margin against that control arm of IV-only treatment, which had an event rate of 6.3%.
So You know, those children in, low-, middle-income countries don't necessarily be-need to be hospitalized for five days in a row for IV antibiotic treatment.
Emily: And I guess it didn't even come close to the 10% non-inferiority, you know, some people might say 10 is wide, blah, blah.
Yeah, But it looks quite safe, right? Like, if we look at the absolute rates of readmission, it's kind of low across all three groups and, yeah.
Angela: They're so similar. Yeah, yeah. but you're absolutely right. The margins are very wide. And these margins, these non-inferiority margins get even wider when you see how few events there are,
and you are asking quite a lot because the primary outcome in this case, it isn't just readmission, it's readmission or death. So you are kind of saying, "Well, it's fine with me if, you know, there are three times more events," meaning potentially death in this arm, and it-- and I'm still gonna call that non-inferior.
But You're right. That's a huge problem that I think all of us trialists have. Like, we simply [00:33:00] cannot afford financially to have really small non-inferiority margins, which clinically would be the only thing that's really acceptable, right? Clinically, you would not like to accept three times more mortality for your patients if your outcome really is going to be mortality, something so hard, you know.
Emily: although the door is interesting because that helps to, give those a hierarchy, meaning, you know, death is worse than readmission, and it helps to power your study. But I think still clinicians have a lot of challenges in interpreting the door result.
So that is a limitation.
Angela: Yeah. It's moving along, DORA, I would say. Like, I think it's really spreading into the more mainstream. but I do agree, it, still is not totally self-evident to everybody. and you're right, the DORA would really nicely nuance these things. but DORA was not used in this one.
So moving to the duration question of this study, every total duration studied, so from four days all the way to seven days, [00:34:00] was non-inferior to what they used as the control arm, which was eight days, and that's again in the PO switch arms. They had one group that went to eight days of follow-on therapy, and everything was essentially the same.
it did not make a difference how many days you were giving, from four to eight. And the good news was, as expected, oral switch shortened total hospital stay by about one day, as we might expect, while each additional allocated day of treatment added about almost one percentage point of antibiotic-related adverse events.
Emily: And it, it jives with our pneumonia duration data. So, you know, if we think about, like, the sum total of evidence for shorter treatment for pneumonia, that aligns, which is always really nice.
Angela: My biggest questions with this trial would be that wide margin of non-inferiority. But like you say, more and more we're trying to find creative ways around it, like with the DORA analysis.
And then I think we just also need to keep in mind that this study applies [00:35:00] well to kids with severe pneumonia, but not to kids who were critically ill, these were not ICU patients. We just have to keep in mind it's, it's still a somewhat select population. But I would say that overall this trial does support early oral switch to just amoxicillin alone and durations of four or five days.
Emily: Yeah. Love it. I'm sold.
Angela: Okay, cool.
Emily: I'm gonna treat 100% of my pediatric population this way. I treat no pediatric patients .
Angela: Oh, no, neither do I. Like, I actually haven't seen a pediatric patient since I worked in Africa, like, 20 years ago. So,, I have nothing to do with pediatrics. But I just, I think it's hugely important and I know we do have pediatric listeners because I hear from them.
Emily: 100%. And I love the trial design. I think it's very fun for us to hear about, novel trial designs, that we can port a- also to the adult world.
Angela: Yeah. you know, right now there's just such a great trend towards adaptive trial platforms, And this was not that. This was not an [00:36:00] adaptive trial platform, but they managed to ask so many questions and answer them within one trial.
it's really, it's a great model for future trials. You know, if you're gonna go in there and do so much work, set up such an infrastructure, get a lot out of it, . It doesn't need to be a full platform that's gonna last for 20 years. , Very few people actually have the resources for that.
If you're gonna do, what we would call a one-off, then do it like this,
Emily: yeah. , Try to answer more than one question.
Angela: So Moving to my second article that I had to choose among so many in the last few months. So truly by chance, my second article is also pediatric, and it is also a duration study.
This one is entitled Individualized Duration of Antibiotic Treatment in Culture-Negative Early-Onset Sepsis in Late Preterm and Term-Born Neonates in Denmark: A Multicenter, Open-Label, Randomised, Controlled Non-Inferiority Trial. You can imagine it's another Lancet study.
This one , was published in Lancet Child & Adolescent Health [00:37:00] quite recently. First author was Emma Machow Carlson, and last authors were Ulrike Nygaard and Tine Brink Henriksen, from the University of Copenhagen. I think you can tell from the way I pronounce their names that I do not know these authors.
I don't have any, disclosures on this trial. And I really appreciate this study because it tries to reduce unnecessary antibiotic use for a condition where there's plenty of overuse. But again, it tries to do it in a very thoughtful and this time very individualized way. So this trial, called DURATION, all caps, was a nationwide non-inferiority randomized trial conducted in all seventeen of Denmark's seventeen neonatal departments, so truly nationwide.
This is why I love Denmark. It enrolled late preterm and term neonates, so anybody being born at least thirty-five weeks gestation. They had to be aged zero to three days, so truly neonates, and they had to have a [00:38:00] birth weight of at least two thousand grams. And to get into the trial, these babies had to have a culture-negative, what they called probable early-onset sepsis.
I'll get into that. So possible and probable early-onset sepsis were defined by a negative blood culture and a clinical decision to continue antibiotics beyond thirty-six to forty-eight hours, which is what they would normally do, because of neonatal clinical signs, maternal risk factors for infection, or both, in combination with an elevated CRP.
Again, CRP over ten. So neonates who already met the usual criteria for stopping antibiotics before forty-eight hours in Denmark were not eligible. So to stop, the criteria were a combination of a low suspicion for infection at antibiotic initiation, only transient clinical sy-symptoms, and a CRP measurement of maximum fifty and decreasing.
and then no symptoms after forty-eight hours once [00:39:00] treatment was started. So those babies weren't allowed in the trial because their antibiotics were gonna be stopped by standard of care anyway. Other exclusion criteria were infections that were clearly gonna need long antibiotic therapy, like osteomyelitis.
So the trial was really targeting infants in that gray zone where you have a clinical suspicion with some signs here and there, exactly the population to whom we may end up giving a full seven-day course of antibiotics out of concern, safety, et cetera.
Emily: Yeah. We see this population of patients all the time.
Yeah.
Angela: I have to say, I really like the way they picked it out. One of the criticisms that they bring up of the trial is that, well, early onset, probable early onset neonatal sepsis isn't really a thing. It's not a real diagnosis. And they're right, but on the other hand, it's exactly a syndrome where you end up giving antibiotics.
Emily: we're stuck treating things that aren't a diagnosis all the time, and so we need some guidance on what to do.
Angela: Yeah. I thought it was thoughtful the way they set it up. So randomization took place thirty-six to forty-eight hours after treatment started, and it was stratified by [00:40:00] two things: site and CRP.
So those two CRP strata were maximum CRP above or below ninety. Infants were randomized one-to-one to either an individualized duration where antibiotics were discontinued after twenty-four hours without clinical signs of infection, as long as CRP was declining to under thirty.
That was the magic number. Once the child got to less than thirty, the antibiotic was stopped. Of course, if there were no, accompanying signs of infection, fever, et cetera. The other arm was the control, and that was standard treatment, which in these centers is officially five to seven days, once they've started the antibiotic in this sort of cloudy population.
Empirical treatment was, and this is so Danish. Empirical treatment was benzylpenicillin and gentamicin, with gentamicin administered to provide three days of coverage. Then an early intravenous to oral switch in clinically stabilized neonates [00:41:00] about thirty-six to forty-eight hours after the initiation of antibiotic treatment.
And that oral antibiotic was mostly amoxicillin. So outcomes. there were two co-primary outcomes. One was hospital readmission for bacterial infection or death within twenty-eight days. That was tested for non-inferiority with a four percent margin. The other co-primary outcome was total antibiotic exposure within twenty-eight days, tested for superiority.
Although I would note that they do not actually define a margin of superiority. They don't give a minimum clinically important difference. they just test it for superiority. Finally, bacterial infection. In that first part of the primary outcome, bacterial infection was defined as at least one clinical symptom of infection and/or CRP above ten and antibiotic treatment for seventy-two hours or more.
So the definition was it was not a microbiologic definition. It did not require microbiologic confirmation. Just keep that in mind. Very real-life study. They did do a secondary [00:42:00] outcome where- they looked at the same endpoints at a hundred days.
and then, of course, they looked at safety outcomes. People can read the article on that one. They, had a target sample size of 488 neonates based on an anticipated risk of infectious readmission of 1.5% in the individualized treatment group and 1% in the standard group.
Upfront, they were ready to accept more readmissions, the price to pay for individualizing and shortening treatment. Here they used a non-inferiority margin of 4%, 'cause they knew, they knew their number of events was gonna be so low that they, weren't gonna try to get away with a 10% margin. 4% is what they chose.
they padded the sample size a bit because they expected non-adherence, in up to 15% of the neonates. and their, their assumptions were actually based on observational Danish data. They've got a lot of it. They note that the 4% margin was defined by consensus among neonatologists, pediatric infectious disease specialists, and clinical microbiologists [00:43:00] balancing clinical relevance with feasibility.
And there I've gotta be honest, I really like their honesty. Like, feasibility means economic feasibility, right?
The results of this trial. Again, you have to love trials from Denmark. They just know everything. They know exactly how many births there were in the whole country, and they have all of the data on those births. So here, okay, this is how the paragraph starts. Between April twenty twenty-two and April twenty twenty-five, a hundred sixty-five thousand one hundred and one live births occurred in Denmark... long story short, they knew that in this time period, there were eight hundred and sixty-eight neonates born with culture negative early-onset sepsis. Of those, eight hundred and eleven were eligible for participation, and of those, they were able to enroll four hundred and ninety-three.
So sixty-one percent of screened were ultimately included and randomized, leading to two hundred and forty-six in the individualized treatment arm [00:44:00] and two hundred and forty-seven in the standard arm. So I think that's pretty good representativeness. We don't often see those percentages. And so who were these babies that got in?
These four hundred and ninety-three babies were almost all term infants, as they were meant to be. Median gestational age was about forty-one weeks. Median birth weight, about three point seven kilograms. And only three percent i-in the individualized group and two percent in the standard group were born before thirty-seven weeks.
Apgar scores were normal, in both arms. About two-thirds had at least one maternal risk factor for infection. And around thirty-eight percent of the neonates had received intrapartum antibiotics already.
So nearly all of the neonates had clinical signs of infections when the antibiotics were started. The Danish are very reasonable when they start antibiotics. mainly they had tachypnea, behavioral changes. About forty percent needed respiratory support, almost always CPAP, while circulatory support was quite rare in this [00:45:00] population.
antibiotics were started at a median of about one day of life. Median peak CRP before randomization was about 73. So these were babies who had some inflammation going on there. The groups were well-balanced, except that the standard group had more boys, 67% versus 57%. So what did they see? Primary outcome.
Median total antibiotic treatment within 28 days was 2.8 in the individualized group, versus 6.8, so almost seven days in the standard group. So that's really, really nice 'cause, you know, you do kinda wonder, "Well, how much is CRP really gonna bring down your actual duration of antibiotics?" And it brought it down a lot.
These babies were dropping their CRPs pretty fast. so there was a difference of four days, between these two arms. Clearly, the individualized strategy achieved superiority with this difference. although again, no minimum clinically important difference was ever actually defined. But sure looks superior, right?
Emily: Sounds clinically important to us, I think, yeah.
Angela: For the [00:46:00] other side of that primary outcome, the individualized strategy was non-inferior for infection-related, related readmissions at 28 days, and per protocol results were consistent with that. The actual numbers were super, super low, and this is where it was so cool to see a 4% non-inferiority margin, but guess what?
There were only two events in the individualized arm of 246 babies within the 28 days. And there was one event in the standard arm at 28 days. At 100 days, there were five events in the individualized arm versus four in the standard arm.
So definitely non-inferiority, but again, really low event rate for a 4% non-inferiority margin. So the authors conclude that a clinically guided individualized strategy can safely cut treatment by four days to a median of about three days, which would lead to a lot of reduction in antibiotic use in this population.
lots of limitations, but overall I think it was a cleverly designed study. [00:47:00] I am not a pediatrician. it will not change my practice.
Emily: But it's quite reassuring to me, though. I love it. Yeah.
Angela: Yeah. I'm intrigued by CRP, . I mean, we ran a whole RCT in adults, looking at CRP to shorten, antibiotic durations, to reduce antibiotic use.
It's used so often here in Europe, but at least where I work, we don't get it routinely. I have to say, the internists particularly are very good at getting it when they have a question, , when they have a clinical question that they are asking,
, What I do as an infectious disease doctor who's called to help out, I will use the CRP because a CRP is so sensitive, and it does come down before fever will. So I will use it essentially to reassure a physician to say, "Hey, we don't need to widen the spectrum.
We don't need to ramp up. Give it a little more time. your patient is really about to turn the corner. You can't see it yet, but I can see it with the CRP." We'll use it in targeted ways like that because it's so sensitive. It talks to you [00:48:00] before anything else is really talking to you.
If you know how to use it, and when to use it, then it's a really cheap, really quick, easy tool. Requires lots of nuance.
Very cool.
Emily: Yeah.
Angela: All right.
Emily: So thanks everybody for listening. We hope you enjoyed the episode, and you're all caught up now on some of the latest research from the last three months. Thank you so much to my co-host, Angela, who is joining us from Geneva. I'm in Montreal. This, quarterly catch-up episode, presented four articles of interest to us that We hope you liked it. Look out for these episodes to come out every three to four months to help you stay in the loop on some of the latest infectious diseases and microbiology science.
If you liked the episode, feel free to let us know by using the like button, share it with a friend or a colleague. It was edited by Katie Hostetler, and the theme music was composed and conducted by Joseph McDade, and the executive producer of Communicable is Angela Huettner. Any published literature we've discussed today can be found in the show notes, and you can subscribe to Communicable wherever you get your [00:49:00] podcasts, or you can find it on ESMID's website for the CMI Comms Journal.
Thanks for listening and helping CMI Comms and ESMID move the conversation in ID and clinical microbiology further along