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 E064 on macrolides
===
[00:00:00]
Marc: 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 Marc Bonten. I'm a professor of molecular epidemiology of infectious diseases and currently vice dean of education at the University Medical Center Utrecht in the Netherlands.
And today I'm joined by my colleague, Emily McDonald, also CMI Communications editor.
Emily: Hi, everybody. Coming to you from Montreal in Canada at McGill University. I'm a professor of medicine in internal medicine, and happy to be here today to co-host this episode. Today's topic is a long-lasting clinical question, and it relates to whether or not we really need to prescribe macrolides in the empiric antibiotic treatment of community-acquired pneumonia.
So more specifically, do [00:01:00] we need to add a macrolide to beta-lactam antibiotics in patients hospitalized with an acute respiratory infection for which possibly the antibiotics that they're taking may already cover them? We're not talking about patients referred from long-term care facilities or patients with severe underlying diseases or pathogens that are unusual that we suspect need coverage.
We're really just addressing the person who's come in with community-acquired pneumonia. We're covering them with a beta-lactam. That more than likely could cover their cause of the infection. Do we need to add a macrolide on top of that? And we're also not addressing long-term treatment with macrolides for immunomodulation today
although it's an old topic, today we have two young investigators to discuss this topic.
We welcome Sylvain Lauter from the University of Manitoba in Canada, and Valentine Schweizer from the University Medical Center in Utrecht, the Netherlands. Sylvain is an infectious disease and critical care physician and clinical scientist who conducts clinical trials in [00:02:00] CABP, and Valentine is a clinical microbiologist and epidemiologist who did his PhD on identifying the optimal empiric antibiotic treatment, very much fitting with the topic of today
So as our listeners know, we usually start these episodes with a get-to-know-you question.
and for this week, our question is, what was the last movie that you saw in the cinema? So I'll go first. The last movie I saw in the cinema was Barbenheimer. So you may remember this was Barbie and Oppenheimer. it was both movies that came out at the same time. It became a very big cultural phenomenon to see both movies.
Emily: quite a difference between the subject matters, of course. but on social media, they created what's called a portmanteau, and that's when you take two words and you combine them together and you make one word from them, and that word was Barbenheimer, and I really liked the idea of it. So I saw those two movies together in the theater, as was a big cultural [00:03:00] phenomenon at the time.
over to you, Marc.
Yeah. For me, it was the movie called Whiplash, and I saw this movie as part of an education activity in our medical curriculum, and the title suggests that this was a, a trauma, movie or whatever, but it was not. It is about a young, very talented musician, a drummer, in a very highbrow, music conservatory in New York and, gets the most, famous, professor who has to teach him, and he is kind of a drill instructor.
Marc: And this movie is about how far do you go in, your ambitions, and how much can you accept? And I think the students choose this, as a reflection of their own future career, but I very much enjoyed the movie, and I did not recognize myself in any part of that movie. So that's the one I saw, Whiplash.
Now we go to, Sylvain.
Sylvain: Yeah, thanks. I'm, actually embarrassed to say that it's been probably over 10 years that I've been to a movie theater, so I, don't really remember the last, movie that I [00:04:00] saw in the theater. I've also had a baby, which makes it very difficult to, get away for an evening out.
But, we did manage to go to a live theater play in New York. We were there for the World Cup, this summer. and we saw "Hamilton," which was fantastic. So, good to see people singing and performing live. It was great.
Emily: And then over to you, Valentine.
Valentijn: Yeah, thanks. I can very much relate to Sylvain. we really used to be big movie enthusiasts, but, we got a daughter about 15 months ago, so, going to the movies regularly has not been, fitting in the program unfortunately. But, the last movie I saw was, I Was a Stranger, which is about, five different people experiencing a traumatic part in their life. and they meet each other's, at a certain point , in the movie. it's very impressive, not for the faint of heart, but, I would highly recommend the movie
Marc: Thank you. Okay, let's get started. first, let's talk about the current guideline recommendations. Valentine, could you [00:05:00] briefly describe them and also, tell us what is the current practice in your hospital for the question that we address today
Valentijn: Yeah.
So in the Netherlands and also in my hospital in Utrecht, we follow, uh, the recommendations of the Dutch national guideline, which was actually published in CMI communication, this month. And in this guideline, the empirical treatment of, adults with, community-acquired pneumonia is based on, disease severity at presentation.
so the underlying philosophy, is that if a patient is, severely ill, the empirical therapy should be, broad-spectrum, and thus cover every potential, pathogen that you could encounter in CAP. and if the patient is not severely ill, there is room to start with a more narrow-spectrum antibiotic and, either perform good, diagnostics so you can escalate or switch to targeted treatments.
Valentijn: or if the patient, deteriorates, then you still have the option to escalate, if needed. the severity that [00:06:00] we use, in the Netherlands, is both the CURB65 score and the PSI score, and these are, used, exchangeably depending on the hospital.
and as, pneumococcus is the most commonly identified pathogen in CAP, in the Netherlands, the empirical treatment in non-severe cases at least focuses mainly on pneumococcus. So, empirically, we start with, either amoxicillin or penicillin. and in these patients, we do not, recommend atypical coverage, so not with a macrolide or not with a, quinolone at all.
And this, recommendation to start with, beta-lactam monotherapy, so amoxicillin or penicillin, is largely based on the, CAP START trial, which is, a Dutch, cluster randomized trial that was published in the New England Journal of Medicine. And in this, trial, which included patients hospitalized with non-severe CAP, not the patients, but the wards were randomized to implement a strategy.
Valentijn: and in this strategy, the empirical [00:07:00] treatments, was adopted for a certain period. and these, strategies could be either beta-lactam monotherapy, beta-lactam with a macrolide combination therapy, or fluoroquinolone monotherapy. And, the fact that we randomized a strategy in a ward and not individual patients means that, deviation from this strategy, by the treating physicians was also possible.
the primary outcome of this, trial was, non-inferiority on ninety days mortality. And the results of this trial actually showed that, beta-lactam for these non-severe CAP patients was non-inferior both to beta-lactam macrolide combination therapy and fluoroquinolone monotherapy.
So therefore, in non-severe CAP patients, we start with amoxicillin or penicillin. And in severe CAP patients, of course, there is less room for error, and we do start with, atypical coverage
Emily: So Sylvain, considering these guideline recommendations that we just [00:08:00] discussed, we're wondering, you know, where do you see equipoise and what's the current practice, for example, in your hospital?
Sylvain: Yeah, great question. I think, hopefully we'll get into where the equipoise is during this discussion, but I think the real question is whether or we need a macrolide or not, and specifically which patient population, may benefit, if any.
so I would generally agree with the approach of Dr. Schweitzer, and we use a very similar approach, I think, in Canada, particularly, in the hospitals where I work in Winnipeg. we follow guidance that's adapted from the Canadian Antibiotic Treatment Guideline, which is supported by the Public Health Agency of Canada and also AMMI Canada, which is our national ID network, and it's largely in line with the ATS and IDSA guidelines, with maybe some, small exceptions.
so we do also stratify by illness severity in hospital when we're determining what antibiotic to use for empiric pneumonia [00:09:00] treatment. we also use a CURB65 score. some hospitals will also use the NEWS 2 score to predict, disease severity or monitor for worsening. and these patients are typically classified as mild, moderate, or severe.
we often will think of mild patients as patients that are admitted to hospital without the need for oxygen. moderate patients on the ward are patients that are admitted to the ward, not critically ill, but they do require oxygen to treat hypoxemia. And severe patients are more the ICU population, where they may require organ support or at high risk for decompensation.
so generally speaking, we would also use a beta-lactam without a macrolide in the more mild population. Typically, we'd use amoxicillin as our beta-lactam alone. In the more moderate population, the guidelines suggest to use a beta-lactam alone. but what I would suggest is that there's a lot of deviation with clinical practice on the [00:10:00] ward, and there's also variability in which clinicians actually adhere to these guidelines or not.
so I find that there is a lot more leniency or tendency to use a macrolide in the more moderate population on the ward. Those are patients that require oxygen or may have physiologic variables that, would be concerning for deterioration. and in the more severe population in the ICU, we would use a beta-lactam, typically in combination with a macrolide.
and the beta-lactams that we generally use in the more moderate or severe populations tend to be third-generation cephalosporins. We acknowledge that there is really a lack of strong evidence that the addition of a macrolide reduces mortality or improves clinical cure however, as mentioned in patients particularly with severe CAP, they may benefit from broader spectrum antibiotics to cover the range of potential pathogens from a community-acquired infection, where there's maybe less time or a shorter window, where effective [00:11:00] antibiotic administration could be a critical determinant of survival.
Sylvain: And so it's, really, more of a feeling than an evidence-driven decision. but, I think more data is needed in this domain
Emily: I'll maybe I can throw in what we do in my hospital, because we're a transplant hospital. it's a bit of an interesting approach, I think.
so for our community-acquired pneumonia who's, let's say, well enough to be admitted to the medical ward, we use a lot of doxycycline instead of macrolides, because of the drug-drug interactions with the macrolides. If we suspect Legionella, then we'll lean more towards using azithromycin. But we do rely a lot on doxycycline, in my hospital.
So I'll just sort of add that nuance that we brought into our local guidelines.
Marc: And is that then doxycycline alone or doxycycline added to a beta-lactam?
Emily: It's added to a beta-lactam. Okay. So we do a lot of amoxicillin-doxy
Marc: So now let's ask a question for which I hope you can give us a very [00:12:00] honest answer.
What proportion of the patients that we described here is the patients that have community-acquired pneumonia, are admitted to a regular ward, how many of these patients, what proportion really receives amoxicillin or penicillin as the guidelines recommend? There's guidelines and there's real life.
Valentin
Valentijn: Yeah, we actually studied this, also in a stewardship implementation trial in the Netherlands. Yeah, the trial is one thing. So if you do a trial like you did the CAPARC study, then it's a protocol and people tend to adhere, although the adherence is far from 100%.
Marc: But outside the study environment, what do the people really do? When I was young and treating patients, it was ceftriaxone that went in at the ER. No amoxicillin, no penicillin. Has that changed?
Emily: I don't think so, no. We use a lot of ceftriaxone
Valentijn: Yeah, [00:13:00] I agree. Also in the Netherlands.
Marc: So we did the studies. We have the guideline, but we're still not practicing that. Is that the conclusion?
Emily: Yes, professor.
Sylvain: I, I think the implementation lags behind what the data suggests. And, I have to say that, I'm not 100% sure that all clinicians know the data behind why we do the things that we do.
And I think, you know, pneumonia being one of the most common reasons for hospital admission around the world, it becomes a knee-jerk reaction. When you see a community-acquired pneumonia, you do what everyone else is doing, and you don't think about it too much. And so, I would say that certainly in our hospitals, following the guidelines or even looking at the guidelines is a rarity.
Marc: Okay. That puts us in a very specific area where we are with data supported by studies with a guideline and a population of physicians that's just not following the guideline. And we have practices in ER where they say [00:14:00] if the patient looks septic, then ceftriaxone or another cephalosporin should be given within the hour.
That's in the surviving sepsis guidelines, I think. And I think everyone with a community-acquired pneumonia, at least most of these patients, will be considered to have sepsis. So is there still a need to have guidelines for community-acquired pneumonia?
Sylvain: It's a good question. I think there's a lot of overlap between those syndromes.
pneumonia is probably the number one driver of sepsis in hospitals. we have sepsis guidelines, we have septic shock guidelines, we have pneumonia guidelines, and a lot of these things have overlap. know, I think it's complicated to tease that out. But, the other thought that I had as you were speaking, Dr.
Sylvain: Banton, is, the guidelines are informed by data, but sometimes the guidelines are informed by data that we wish was stronger, we wish was more robust data. And I think particularly in the field of antibiotics, for empiric treatment of pneumonia, I think there's [00:15:00] room for us to generate more robust data that could maybe, further inform guidelines with stronger stance and higher level of certainty behind those guidelines.
And that may be something that's needed to influence the clinicians to follow the guidelines.
Valentijn: Yeah. So I, just, double-checked, in the CAP-PACT trial that we did in the Netherlands, we actually measured, how many patients were prescribed narrow spectrum antibiotics, before we introduced the intervention. So this is not, biased by observing, um, having effect there.
And, about, uh, 30% of the patients, that were admitted were actually treated with narrow spectrum antibiotics, so either penicillin or amoxicillin. I do agree pro-probably one of the main reasons why physicians start, uh, with broader spectrum, beta-lactams is because the diagnosis, when the patient is admitted, is not so clear.
So it's a septic patient. It can either be a pneumonia or a urinary tract infection.
Marc: Yeah. So coming back [00:16:00] to Sylvain's point on we need better data to convince clinicians to adhere to the guideline, and the guideline would be to use, a, monotherapy with a beta-lactam antibiotic as the, let's say, the standard of care.
what would be the best study design to deliver that, convincing evidence? And I think, let's start with Valentine and then Sylvain.
Valentijn: Yeah. So I think in general, a randomized approach is preferred, compared to observational studies. And I think the, recent results that we saw from, oseltamivir is very, illustrative because also in this case, you had a lot of observational studies showing that there was some mortality benefits, of oseltamivir in flu patients.
But now the randomized REMAP-CAP study shows, that if you randomize, you actually see potential harm of the therapy. So the, confounding by indication that is present in observational studies is very difficult to correct for. you never know how [00:17:00] much, residual confounding is still present after you try to correct for confounding by indication.
and even in novel observational, designs like target trial emulation, there still is the problem of confounding by indication. So in the CAP-START trial, we chose a cluster randomized design, because you want to investigate the empirical therapy, and you want to start with the randomized therapy as soon as possible.
and especially in studies where you want to study the empirical therapy, any antibiotics that is, given before randomization, they, dilute your treatment effects by quite a large margin. And I think this is, especially the case if you have, diseases like community-acquired pneumonia, where, there are a lot of, trials that show that you might only need three or five days in total.
So one dose of ceftriaxone at the emergency department, um, you treat the patient for one day of the total three days that are required. so these, pre-randomization [00:18:00] antibiotics, they have a really strong bias in most, individual randomized trials. so that is the reason why we opted for, a cluster randomized trial.
however, like I was saying earlier, if you randomize, a strategy, then it is possible for physicians to deviate from the strategy. And this deviation, this non-adherence to the treatment that you're actually investigating, that also results in a dilution. So all these patients that are treated differently, if these patients are similar in your different, treatment strategies, then these patients will also result in a bias to non-inferiority.
so that's one of the major disadvantages of cluster randomization. I think, looking at what kind of research design will eventually be the most, optimal for studying empiric antibiotic therapy, they have to be fast in randomizing patients. So looking [00:19:00] at, novel research designs where, uh, randomization could somehow be integrated in the electronic patient files, like what's done with the ACORN trial in the US where they studied, whether patients empirically needed anti-Pseudomonas therapy.
here they were able to randomize patients, very quickly, without the need for individual informed consents, because the therapies that were evaluated were considered low-risk therapies, and there was a lot of, variation in practice, already, and they were not considered to be, any issues regarding safety.
and I think using, a similar research design in community-acquired pneumonia, reducing the influence of pre-randomization antibiotics, that is something that is actually also needed for CAP.
Marc: So Sylvain, what, is your take on the effect of this pre-randomization antibiotic exposure?
So people that are, treated in the Netherlands, for example, by their GP, [00:20:00] and after a few days they reach the hospital, then actually our possibility to do something starts at the hospital entrance. So we're talking about that, pre-antibiotic use in the hospital before the patient is in the trial
Sylvain: Yeah, I think it's very important and it becomes messy very, very quickly.
patients that present to hospital with pneumonia may often have had symptoms for a couple of days before they come to the hospital. They might have been seen in a clinic like a walk-in clinic or a doctor, and they're often prescribed oral antibiotics in the outpatient setting in the days leading up to their hospital admission.
Sylvain: So that is something that's out of our control in clinical trials. it should be tracked and monitored and accounted for in the analyses, but clouds the picture. Then you have the, time window from presentation to hospital to randomization, and hopefully that's a short interval. but you know, with the surviving sepsis guidelines that we've talked about [00:21:00] and the culture of treating patients that are presenting with infections, there's a tendency to give rapid antibiotics in emergency departments.
In Canada, we even have, a few provinces that have adopted nurse-administered antibiotics that are being given in waiting rooms of emergency rooms, because of the concern over time delays in sepsis patients. And so these are antibiotics that are administered in the waiting room before they're even seen by emergency room physicians.
So further increasing our challenges of getting to these patients and randomizing them early I agree that we absolutely need randomized data and cluster randomized trials are a great way to randomize a strategy. they have their limitations, as Dr. Schweitzer has mentioned. but the CAP START trial is a great example of how there can be success in generating very important data, using a cluster randomized design.
I do think that having individual patient-level randomizations is [00:22:00] important to increase power, to have granular data, but getting to these patients really early, is absolutely necessary. So I think we can do this, individual patient-level randomization in an empiric antibiotic trial. but the design really needs to be novel.
Like it needs to be pragmatic, it needs to mirror like standard care procedures in hospital. the data and the barriers to randomize has to be very low. waiver of consent or deferred consent, I think we could argue that that would be justified and ethical in this scenario where we really don't know there's a ton of equipoise of which antibiotic to use.
We're already using them in various different ways across different centers and we really need to generate this data and the time is an emergency, to get these antibiotics into patients. So I think a deferred consent model or even waived consent model might be warranted.
we have done some work to incorporate, embedding eligibility criteria into EMR [00:23:00] systems, and so that's having the computer identify patients with symptoms, identify patients that would meet eligibility criteria, so that a flag is raised and clinicians are notified that this person is eligible for a clinical trial.
We've also done some work with our site networks and our current clinical trials where we are very active in the emergency departments. So historically, our trials teams have been embedded in the ICU or on a ward, but we're really trying to break those geographic barriers and move throughout the hospital as the patient moves through the hospital journey.
And so really capturing patients early with the trial teams being down in emerge when the patient is first presenting would allow us to tighten that window and hopefully avoid or at least minimize the dilution. So I don't think it will ever be perfect, but I think with newer study designs, newer analysis methods, also primary outcome selection, I think we'll be able to generate some [00:24:00] really important patient-level data, using trial methods
Marc: Yeah.
Just, to stick on the contamination by pre-randomization antibiotic exposure. I can tell you that was one of the reasons why we came up with a cluster randomized design, as that was the only way to circumvent that problem. Gave us other problems in the cluster randomized design, but this was an important one to circumvent.
What would you say about the validity of studies addressing this research question in which the in-hospital pre-randomization antibiotic exposure is not controlled? So there is exposure. let's say, 24 hours on average, that a patient is randomized and the patient has received maybe a long-acting antibiotic like cefriaxone in the emergency room.
What would you say or what do you dare to say about the validity of a study that is then comparing the effectiveness yes or no, of, a macrolide?
Sylvain.
Sylvain: it's a very hard question. I don't know the answer, but my gut feeling [00:25:00] says that it does make a big difference. I think that, the exposure of a single dose of a long-acting antibiotic may influence treatment. there's good data about duration of therapy in community-acquired pneumonia, and typically shorter tends to be better.
and we really don't know the optimal duration. Like maybe even the guidelines would suggest five days is probably adequate for a treatment course for pneumonia. But it's possible that even a shorter course of therapy may be beneficial. So how much antibiotic does the patient really need for community-acquired pneumonia?
Maybe a single dose, maybe two, three days' worth of antibiotics is enough to sterilize and eliminate the pathogen. and then what you're left with is really the host response and the inflammatory response, and th-those are really the main pathophysiologic mechanisms that drive symptoms and illness severity.
so I think it does make a difference what patients are being exposed to in emergency departments.
Emily: Yeah, I guess you could stratify, or you could hope that the randomization, you [00:26:00] know, takes care of those, differences in the baseline population. If we take a step back, part of the problem-- And Sylvain and I have run a couple of pneumonia trials now together.
It's actually a difficult diagnosis to make. It's not like a bacteremia trial. you often don't have the microbiology, and so many of the patients in the trial to start with won't have a bacterial pneumonia. You know, we'll be randomizing people in that trial who have a viral pneumonia, or who have aspirated.
so it's very tricky to run pneumonia trials, I think, 'cause sometimes we don't even know that they have the diagnosis that we're putting them in the trial to treat.
Marc: But then the question is, who should be in the trial? Should it be the patients with a bacterologically documented cause of infection and a radiography supporting pneumonia, or should it be all the patients that we treat as if they had pneumonia?
Valentin, what do you think?
Valentijn: Yeah, that's a very difficult question to, [00:27:00] answer. if you have very strict, selection criteria for the patients that you will include in your trial, then obviously you're gonna have problem generalizing these results to your clinical practice, right?
And we know from studies in community acquired pneumonia, also looking at, whether patients had, an infiltrate on a chest, uh, radiography, for instance. That if you do low dose CT scanning in these patients, then, in a lot of these patients you don't find it on chest radiography, but you do find an infiltrate on low dose CT scan.
Meaning that if you start, screening and only including patients with a chest A and an infiltrate, then you will probably miss also an important part of the patient population that you are actually treating
So we're now halfway there,, and we have only discussed the problems at the entrance of the study. Now let's talk about the outcome of the study. So what, should be in an [00:28:00] antibiotic trial of community-acquired pneumonia where we compare a beta-lactam to a beta-lactam with macrolides, what would be the best outcome, the primary outcome of that study?
Sylvain: Yes. I think that's a great segue because it actually-- choosing the outcome ties in nicely to the contamination discussion that we were talking about in the beginning. I think that, you know, if there's going to be contamination in a trial, and the trial is designed as a non-inferiority trial, it's likely to bias towards no effect.
and so, you know, teasing out that effect, is gonna be influenced by the exposures patients get before hospital and when they first present. I think that, showing non-inferiority between these different regimens is also not likely to influence practice because I don't think that clinicians are scared enough of the potential harms of overexposing people to antibiotics.
And [00:29:00] so, you know, showing that a beta-lactam alone is non-inferior to a beta-lactam plus a macrolide, I don't think is gonna be a game-changing practice finding that all clinicians will read the paper and go, "I'd have to change my practice and stop." I think what we have to do is show that, using a macrolide is potentially harmful to patients.
and by doing so for antibiotic trials, I really think you need to structure your outcome into an ordinal or hierarchical model that shows, both the clinical efficacy and clinical response, but also the potential harms of over-antibiotic exposure and antibiotic use. So there are newer, clinical trial outcomes such as the DOOR outcome, which is a desirability of outcome ratings that are becoming more common and popular in infectious disease research and in other fields of medicine.
And it's a framework that reflects the overall clinical experience of each patients rather than analyzing efficacy and side effects or adverse events [00:30:00] separately. it's particularly useful for antibiotic strategies where the best strategy is to achieve recovery while minimizing harms and minimizing overexposures.
so for example, you might structure an ordinal outcome being best outcome being clinical recovery with no adverse event, clinical recovery with an adverse event, clinical failure that requires either escalation or antibiotic change, and then bad clinical outcomes like ICU admission, organ failures, and lastly, death.
And you can see that each patient may fit into a different box in a different level, and we would compare how those different levels change in terms of a, more desirable outcome for patients based on their randomized allocation. And so I think by using, maybe more novel primary outcome designs, we might be able to better address, which is the best, 'cause that's really what we're trying to get to, is what is the best antibiotic.
Best might encompass clinical efficacy but also tolerance and [00:31:00] lack of excess antibiotic exposure.
Marc: Valentine, do you agree on this?
Valentijn: Yeah, I, partly agree. but I also think that one of the most important things is to convince physicians that, more narrow treatment regimen that you're trying to promote, or that you're prescribing is also actually safe for patients.
I think one of the issues in stewardship, in general is that you're, trying to balance, possible detrimental outcomes in patients compared to long-term, outcomes, of the emergence of resistance. So as we are trying to promote prescribing, more narrow spectrum antibiotics, I think, physicians are fine with this, as long as they are convinced that the narrower spectrum antibiotics is actually safe for patients, and we're not harming patients.
and picking, composite endpoints where we both combine, clinical outcomes, with, possible superiority outcomes like reducing the spectrum [00:32:00] of the antibiotic or reducing duration of the antibiotics, can be difficult if you get a mixed, signal, that combines these both co-primary outcomes.
So, I think , the main point is to first demonstrate that it's safe, and if we demonstrate then it's safe, then, you can demonstrate, additional benefits by having a more narrow spectrum antibiotic or having shorter treatment duration.
Emily: Yeah, I mean, door actually lends itself really well to that, because you could, for example, break ties between patients with something like a radar, where you score higher, if you have a shorter duration of antibiotics or less exposure to antibiotics
Sylvain: Another comment that I might add to, antibiotic trials is that, I think we've kind of overlooked the economic and environmental impact of some of the antibiotic selection choices that we make. I think that these are becoming more and more common to embed within clinical trials to understand an economic [00:33:00] evaluation of the intervention and also the environmental impact of that intervention.
you know, for example, IVs and supplies and other, supply chain kind of issues. And so maybe oral pills would be just as good and, easily tolerated, have less economic and environmental impacts. given how common pneumonia is, it's probably one of the most common decision that clinicians in hospital make is what antibiotic should I use to treat a pneumonia patient?
Sylvain: So the scale of this decision, even very small incremental changes in an economic saving or an environmental, , protection, , could lead to large scale improvements in budgets and environmental footprint because of how common this decision is. And so I think that's another important thing to embed within future trials to understand this concept.
Emily: Yeah, I think that's a very good point.
Marc: So coming back to this trajectory that Sylvain was describing, a vision of we should add a macrolide to a beta-lactam based [00:34:00] on, and I use that word very low quality observational studies, polluted with confounding by indication.
That's where this whole discussion comes from. Then we have had some trials that try to randomize and ended up with cluster randomization and gave a signal that it is safe not to give a macrolide. But our future is that we have randomized studies that demonstrate actually that the macrolide is harmful to add.
At least that's what Sylvain was describing. Now, this future trial should take care of this contamination issue before the randomization. I'm now at the point after having done the CAPP Start study and the CAPP PAC study and the CAPP Next study, all with a form of cluster randomization, that the cluster randomization is not the answer.
It should be individually randomized, but then you need to take care of that contamination issue. And we [00:35:00] have been very unsuccessful in doing that because of the ethical regulations that actually forbid us to do so. I want to hear from both of you, what would be needed to bring our society at that stage where we can really do the comparison to give an answer that our patients will really benefit from?
Valentin, maybe you can say something about what we encountered when we proposed to treat patients according to protocol randomized before they could give an informed consent
Valentijn: Yes. So, like I was saying earlier, we were inspired also in our research group by the ACORN trial, Th-this was a trial performed in the USA where they were able to randomize patients within the electronic patient files, without having the need for asking individual consent because these were considered, all to be, low-risk interventions that they were evaluating, and it was considered equipoise there.
And we asked our Dutch, uh, [00:36:00] National Ethics Committee, under what kind of circumstances, would we be able to do such a study in the Netherlands or Europe? And they, had no clear answer for this, unfortunately, because currently we don't have any ethical framework that supports, these kind of, waiver of informed consent, in individual patient randomization
Marc: How's that in Canada?
Sylvain: I mean, in, Canada, we have been able to, use deferred consent and even in some situations, waived consent depending on the study question, the study population, and of course, this would be, backed up by ethics. and also I think the other really important part is to have patient engagement.
people with lived experience and, members of the population express, the importance of doing this and the acceptability, from their lens in doing these types of research. as you say, there's been precedent for using deferred consent models in similar types of research.
Sylvain: certainly we used a lot of deferred [00:37:00] consent, in the pandemic, where there were a lot of uncertainties, and I think it was, widely accepted that we need to generate data, and we don't know how to treat this new infectious disease. lots of deferred consent, is precedent in the intensive care unit where patients are critically ill and often unconscious and may not have family or next of kin readily available.
And so I, think there's precedent here to, do this. And I think that these are-- in general, they're low-risk studies. We're using standard of care antibiotics that are commonly prescribed in any which way, and that's really random care. And in a sense, we may be providing safer care by enrolling patients in randomized trials where we're monitoring whether that decision was a good one or not.
And we're generating data so that we don't participate in random clinical care, but we're, practicing evidence-informed care. And so I think the argument could be made that there's potential safety benefits from enrolling in a clinical trial. and the usual [00:38:00] things like exposure to data or privacy issues, can be dealt with when there's time to have an informed discussion with the patient after the randomization has occurred
That's clearly a different situation than in our setting
Valentijn: Yeah, I completely agree that we should move from random care to randomized care. and I think the SNAP trial is also a great example of this, right? Because we had been using flucloxacillin, uh, for a very long time in standard care.
and now from the SNAP trial results, we see that we were actually, harming patients. so I think if we want to move the field of infectious diseases forwards, we need to look very critically at what kind of therapies we actually have strong evidence for in our patient population. And if we don't, we need to do, good randomized studies.
But even the SNAP trial and the REMAP-CAP trial do not, let's say, control for the pre-randomization exposure to something that influences the effect of the [00:39:00] intervention. it still has this time window. we really need something like the, ACORN trial?
Marc: in the Netherlands, the last failure trial that we had was that we compared a cephalosporin versus a cephalosporin plus an aminoglycoside for the empirical treatment of presumed gram-negative sepsis. And we asked for permission to do this, uh, because there's clear equipoise.
Both strategies are widely used. The safety of the strategies is well known, and there is a need to, evaluate their effectiveness especially. We were not allowed to randomize patients before informed consent. We could use deferred consent, but only to start collecting data before they had approved, so not to randomize the antibiotics.
So, Sylvain, what you're describing in your situation is that you can really randomize between antibiotic A and antibiotic B and then have a deferred consent, and if the patient says, "Actually, I don't consent," they still were treated with something that was randomized. Is that the [00:40:00] case?
Sylvain: Yeah, totally. In Canada, our deferred consent models, we are permitted to randomize and to use the treatment allocation before individual patient consent has been made.
we always, make our best efforts to provide consent afterwards, and as soon as the patient is able to do so. if the patient decides that they no longer wanna participate in the trial, then they could stop the study intervention if it's still ongoing and, the decision would be up to the clinician, to decide, how best to treat them.
Sylvain: and they can also opt to remove their data from the trial.
Marc: Valentine, you must move to Canada to do your trial. Yeah, definitely.
Just pick the right province. It's not the same all across the country. So there is precedent, but you end up, kind of negotiating, maybe 10 times over with each different province and definitely it's still a challenging argument to make
Yeah, Canada is a vast place, and as Emily mentions, there's different, ethics boards and committees throughout each province, so it makes, [00:41:00] activations of sites throughout our country, quite challenging and onerous to have these discussions over and over again.
there are moves, recently to do a unified REB, where there's a single, REB process for the whole country. but those procedures are still in the works.
Marc: So we, have this research question that we think is important, and there are all kinds of hurdles that need to be taken to come to a scientific valid answer on the safety and efficacy of the question, so of the adding macrolides to beta-lactam antibiotics in this patient population.
What about the so what question? What now if that for this indication, we cannot deliver the evidence to give the strong recommendation that it should be monotherapy.
Marc: So why not then give them all combination therapy Who cares?
Emily: I think our oath cares. Didn't we, say something about not doing that? First, do no harm. [00:42:00] So if we are doing something and we don't have evidence of benefit, and we do know that one out of every X patients will be harmed, I would argue that it is our ethical professional obligation, to continue to pursue the question, and we shouldn't just give the antibiotic to everybody.
Marc: But there is a significant part of our community, scientific community says, "If you're not giving meropenem, you're harming patients."
Sylvain: Yeah, I think, I have a hard time divorcing this research question completely. Like I think I would still pursue, and you're right that, you know, it may be very challenging to do randomized trials as we've talked about in this podcast.
But, I think that we're at a breaking point with trial designs now where like we can do this. And so I think with novel trial designs, we should continue to push forward and, break through barriers. so not to give up. But I would also say that, even if we step away from a clinical trial perspective, there are other research methods that we can learn more about [00:43:00] this topic.
and what I mean by that is particularly translational science. I think there's a lack of understanding, physiologically what our treatments do to patients. And so, doing things like observational studies where we collect biospecimens from patients, looking at microbiome studies, looking at host inflammatory response, looking at, end organ markers of damage and, those types of studies they might gain our understanding of what the treatments are doing to patients and might help us understand the potential harms and also maybe even the potential benefits of using different antibiotic strategies.
Sylvain: And so we let clinicians do whatever they wanna do, but we still have ways of collecting data and observing, the impact of those treatment decisions.
Valentijn: I still think it needs to be investigated in a, valid manner.
Sylvain: That's another great point is that I think, the reality of how we perform clinical care is that we don't do any one thing in isolation [00:44:00] the way that we do in clinical trials. And so, for example, a patient with pneumonia might be treated with different IV fluid compositions, different amounts of IV fluids, different oxygen levels and targets, corticosteroids versus no corticosteroids, protective medications like some people would give a PPI for patients that are on corticosteroids to minimize the harms of that intervention.
so we do a lot of things at once when we provide care to patients and, hopefully with the randomization schedule that washes out in terms of even distribution of those, things. But, clinical care is messy, and I think there's a lot to learn about how best to do this. and it may be that, you know, if corticosteroids becomes the standard, I'm not so sure that it's being adopted so much in that more moderate or mild population at the moment.
but as corticosteroids becomes adopted with the emerging evidence, it may be less important to have the anti-inflammatory effects of an adjunctive treatment like azithro or doxy. And so, [00:45:00] understanding those co-interventions I think is also really important in this discussion.
Marc: We talked about the design, we talked about how, when and how patients should enter the trial, and we talked about the outcomes and the ethics Any other points that need to be addressed
Sylvain: I was just gonna make a small plug just to say that, we are starting to try to get to some of these answers with some of our trial designs.
And Dr. McDonald and I, along with a very large team, have developed a platform trial called the GUARD-Resp trial. We have, over 60 sites now activated in Canada, Brazil, the United States, Mexico, and also Colombia. And we're bringing on collaborators in Australia and New Zealand and, and other places in the world.
And so far, the treatments that we've been studying have been host-directed therapies, not so much antibiotic-driven therapies. but we just got some local funding to do a pilot study of empiric [00:46:00] antibiotics in Canada at 10 sites in Canada. And the reason for a pilot study is exactly for the reasons that we mentioned is, can we do this on an individual patient level?
What are the challenges going to be? What's gonna be acceptable for patients and their families? we're planning to enroll around 150 patients. And we're still really early in the kind of, you know, hashing out the details of the study design, the protocol, the steering committee.
and we hope to be able to learn something from that pilot trial in Canada and bring that to other sites around the world. so just an open invitation for those that are listening, and those on the call that if there's people that wanna collaborate in this space, we're a very collaborative group and we'd love to hear from you and, have your input.
Great. what are the host-directed therapies that you're investigating?
Sylvain: Yeah. currently we're enrolling in a domain called Attack CAP. It's looking at therapeutic heparin, to treat the host response in pneumonia. we did this trial during the [00:47:00] pandemic that we published in the New England Journal, and we showed that patients on the ward where, you know, in COVID, where there's a lot of inflammation, thrombosis activation, a lot of clots, patients that were on the ward benefited from, a therapeutic dose heparin.
they had less ICU admissions and increased survival, and they got home sooner. versus in the ICU setting, heparin didn't work in that group. In fact, there was potential, towards harm in the group that was already critically ill on organ support. And so really kind of emphasized for us the need to prevent critical illness and act early on the ward before organs have failed and, and have landed you in the ICU.
so we've translated that question now to all pneumonias. So we've enrolled o-over eight hundred and fifty patients so far across our sites to the heparin domain. so far, the DSMB has met. It, it appears to be safe in terms of the intervention. and we hope to answer the heparin question soon in community-acquired pneumonia.
And we also got some [00:48:00] funding to do dexamethasone. It's a placebo-controlled dexamethasone trial in the moderate pneumonia state. and we have funding to do the full trial. We expect about four thousand patients, . We've just submitted for regulatory approvals, and we'll be launching that through our sites this fall.
we've collaborated with Dr. McDonald and others, to look at pathogen-specific questions, and Dr. McDonald can speak to that, particularly PJP pneumonia. And we're also interested in looking at things like, oseltamivir and antivirals for influenza that we're seeking funding for at the moment and collaborating with, our partners at Remap CAP, which just pres-presented some very interesting findings of potential harm with, oseltamivir in the critically ill population.
Sylvain: So another question that's urgently needed to be addressed on the wards in the pneumonia space.
Marc: Great. So we were co-authors on the heparin papers because REMAP-CAP was part of that as well.
Sylvain: Yeah. we're big fans of the multi-platform approach. [00:49:00] with the emergence now of platform trials around the world, people have different platforms and different site networks, and I think that that's great.
But if we can all work together and federate our data and pool our data, it just means that we'll be able to answer these important questions a lot quicker.
Marc: And those were three platforms together, huh? Three platform trials.
Yeah. ACTIVE4A in the United States, REMAP-CAP around the world- Yeah ... and, and the ATTACC network- Yeah
Sylvain: which was mostly a ward-based network.
Marc: Yeah. All right. Thank you so much for the conversation, and thank you for listening to Communicable, the CMI Comms podcast. This episode was edited by Katie Hostetter. Theme music was composed and conducted by Joseph McDade
And any published literature we have discussed today can be found in the show notes. You can subscribe to Communicable on Spotify, Apple, wherever you get your podcasts, or you can find it on ESCMID's website for the CMI Comms Journal. Thank you for listening and helping CMI [00:50:00] Comms and ESCMID move the conversation in ID and clinical microbiology further along.