How PCT biomarkers can guide antibiotic treatment for sepsis

This transcript was generated using an AI-based transcription tool and may contain errors. It reflects the spoken content of the recorded session and has been formatted for clarity.

Great. Recording is underway. So, as I say, this is the first of our webinar series on introducing the Stewardship Continuum Spotlight on Sepsis and AMR. This session is going to be on biomarker guided antibiotic duration and implementation. And we will have two speakers. So I'm going to be your host today I'm Sheila Money.

I'm part of the medical affairs team in the UK and Ireland and our two speakers who I will introduce will be Professor Paul Dark who's the chief investigator on the Adopt Sepsis trial, and also Jonathan Clayton, who's the national laboratory lead for the adaptive trial.

As I say, this is the first run of the event in our three part series. Today, we're going to look at the Adopt Sepsis trial and the findings of the trial, which was recently published in Jama. We're going to explore the practical aspects and challenges of integration of the pro calcitonin assay into the laboratory and look in detail at the results of the trial.

Our first speaker is going to be Professor Paul Dark. He's the chief investigator of the trial. Paul is professor in critical care and senior investigator and vice dean for health and care partnerships and Faculty of Biology, medicine and health at the University of Manchester in the UK. And Professor Dark will share the pivotal findings of the trial and ask whether the close tracking of biomarkers like therapy and peaked can help guide clinicians in significantly shortening antibiotic courses without compromising safety and looking at it in the specific cohort that was addressed in the trial.

We will then have Jonathan. Jonathan Clayton is a consultant clinical scientist in biochemistry with co applicant and national lead for the Adaptive Optics trial for the laboratory, and is currently at the Lancashire Teaching Hospital NHS Foundation Trust. All of them will offer an in-depth perspective of the laboratory's role in the adaptive trial, outlining the practical aspects and the challenges of implementing the CT assay, and all of the implications for setting it up in the clinical trial setting and also shedding light on implementing it into a routine workflow and the diagnostic strategies.

So I'm going to hand over to Paul. It's going to talk to us about the trial.

Sheila, thank you so much. And what a fantastic opportunity today to help with our dissemination of trial results. I'm here from Manchester. There's a beautiful sunny day and I'm here representing a very large trial investigator and collaborator group. So this is about team science. And I'm particularly pleased that Jonathan's here with me today to talk about not only the clinical aspects of the study, but also the laboratory aspect.

So this was a very firm partnership between clinical practice and our clinical service laboratories. So that's a great opportunity. And I thank Roche for this opportunity. So I'm going to try and cover in my short talk for the next set of 20 minutes or so. What was the need for the trial? I think that's critically important.

The background to why we did this trial, how do we design it to be conducted in routine NHS care or routine patient care in the UK? What are the results? Some of you will know those already, but if you don't, I'll go through those and what we think those results mean. And then I guess the real challenge here.

Is it going to change your practice. And we'll talk a little bit about where we're up to in the UK around routine practice. So I guess any research question needs to start with understanding evidence gaps. And we were very fortunate in the UK that nearly ten years ago, because of challenges in sepsis care pathways, the government had instructed our advisory group called the National Institute for Health and Care Excellence to have a look at the evidence around certain aspects of sepsis care.

And one thing that was prioritized was to think about how diagnostics, and particularly host inflammatory diagnostics, could help with decisions about the use of antimicrobials in sepsis care pathways. And if you want to read that, it still exists in open access online, and it was called the diagnostic guidance DG 18. And that specifically covered host response. And you will see in that guidance that it was very quickly around PCT and CRP as being the two most researched post inflammatory markers in this arena.

And in fact, the evidence for precursor Tonin looked promising. This was ten years ago, but not sufficient evidence to warrant introduction of routine PCT monitoring into care, specifically around the aspect of when to stop antibiotics. So there was a research recommendation from nice at that point to link this with international guidance. And this is our latest guidance, soon to be updated, I believe, for the management of sepsis patients.

This is updated in 2021. And you will know that in that guidance precursor Tonin is suggested as something we could consider alongside patient care for patient monitoring and review daily to make decisions about when to stop antibiotics. But that's quite a weak recommendation. And the evidence base that underpins that is acknowledged to be of relatively low level, even though there are quite a lot of very well conducted trials.

That's the view of that guidance. And there's very little actually on CRP. So there was our sort of starting position from this trial. Now what happens in the UK is that our Parliament and our government make decisions on prioritizing research questions and research funding, and that's called a commissioning research brief. So about eight years ago they produced this statement and challenged us as a community of practitioners to address whether daily CRP and PCT monitoring safely allowed the reduction of the duration of antibiotic therapy and hospitalized patients.

And they specified that they wish to see a definitive three arm randomized controlled trial. So as a community, we have to respond to this. We can obviously produce our own questions, but it's critical that we're response to government directive to show as a community we can deliver research that is meaningful for policy. So that's the motivation here was responding to a government request.

And we have to go very closely to the brief. So in short, I was very fortunate to lead a large team that was successful in winning that commission. And I'm going to tell you a little bit about how we designed the trial. So this is my ICU at the Salford Royal Hospital in Greater Manchester, bed 26, as you can see above the bed.

And we thought very carefully about the kind of patients in our ICU with sepsis that might benefit from interventions using biomarkers to try and limit antibiotic duration. And so when we design this trial, we decided to focus on adult patients above 18 years of age, those patients will be offered the research if they were within 24 hours of having their first dose of antibiotics for sepsis.

And of course, that relates to the kinetics of these markers around capturing the early peaks of these markers. So that was really important as an inclusion criterion. We asked clinicians to judge whether that patient was likely to be in hospital receiving systemic antibiotics in 72 hours time. So we were thinking about a relatively sick group of patients, but left the clinicians to decide whether that was the case.

We didn't think that interventions would be or patients wouldn't benefit from the interventions that we'd planned if they were likely to stay less than 72 hours. And of course, that judgment can be quite difficult to make. As you know, there was some important exclusion criteria in this study around, and you can read them in the paper that I will share later on, but specifically around patients that we didn't think would benefit from the intervention.

So patients that were mandated to have long term antibiotic therapy for very difficult to treat infections, for instance, in bone or heart, with the kind of things we were thinking about, as is often the case of these studies, hematology patients that were severely neutropenia were also excluded. Again, because of the anxiety around a willingness to change practice.

And finally, this study went on through pandemic. And of course, as you know, some very potent anti-inflammatory agents repurposed from rheumatology became very important in Covid patients in ICU that we showed in trials in the UK and in international trials. And those, of course, are very potent at changing CRP and PCT profiles. And so we had to make sure that we didn't include those patients.

So there's very little information in this study about patients with Covid exposed to those agents. In terms of the intervention, we tried to keep this quite simple, that every patient in the study would receive an additional blood sample every day. There wasn't a particular timing to that as long as it was conducted within the 24 hour period and blocks of 24 hour periods.

That additional blood sample was sent to the laboratory. And Jonathan is going to talk more about that. And then the laboratory analyzed relating to the arm of the trial that the patient was in. And I'll talk a little bit more about that in a moment. And the blood samples continued until antibiotics for sepsis were discontinued. So as soon as the decision was made to stop antibiotics, the additional blood sample was finished and the patient went into a follow up phase up to 28 days, assuming the antibiotics had been stopped within that time frame.

And the protocol that I'm going to describe was not reintroduced if antibiotics were reintroduced. So there's a clear intervention phase and then a follow up phase.

We did have a very clear view about the biomarker levels that we were interested in adopting in this study, and it came largely out of the international literature, particularly that nice review. I would say that the evidence underpinning the levels for calcitonin was much firmer than CRP, but there is an evidence based around CRP levels and we adopted that.

I'm not going to go through this in detail, and you can read it in the paper. And Jonathan is going to talk about this a little bit more. But at least you know that in the laboratory we had very fixed approaches to this. Now, what you can see from this diagram is that from the outset, we decided that we were not going to return results to service, but we were going to return decision support.

So information about whether the patient should know that we strongly support stopping antibiotics, that the protocol supports stopping antibiotics or the protocol supports the usual care. And you'll see why we did that in the next slide. Sorry in the next couple of slides actually. So before I move on to the protocol, we had a very significant challenge to overcome, which Jonathan will talk about a lot more.

But there's my bed 26 on one side of the slide, and there's Jonathan's biochemistry lab on the other side. And we obviously have a very high quality digital connectivity. We have pod systems to send blood samples, and the labs know which patients they're measuring on. And we know the results of those samples very quickly. So one of our big challenges in this study was to break that communication, because we wanted to develop an intervention concealment strategy, where clinical service didn't know what laboratory tests were being requested, and that the laboratory service did not know who the blood sample was from.

And that allowed us to break this connectivity across, as you will see, 42 sites across the United Kingdom. And the way we did that was to work with our clinical trials unit at the Warwick Clinical Trials Unit, who built a web based black box where as a patient was recruited into the study and the blood sample was taken, that blood sample was only identified by a study number.

And then when the blood sample was received in the laboratories, the laboratories could find out from that study number what tests the patient had been assigned to. And that allowed us to break that connectivity. And so there was no way of reporting blood samples or advice back into the electronic system. But in fact, the laboratory service entered the results back into the black box, the web based system.

And that produced written advice to the stewardship team in clinical practice. And it's worthwhile saying that regular stewardship is a commissioned service in the UK in most units as part of standard NHS service in the intensive care. So these decision support pieces of advice were going into a daily process around making decisions about antibiotic use.

So very important about the standard care in this trial. So the next thing to say is to talk about our primary outcome, which I think is important to focus on. So in a trial like this and trials that have been conducted very well in the past, it's often the case that the primary outcome is about measuring the duration of antibiotics given for the initial sepsis period, and we thought long and hard about this and just wondered whether there was an opportunity in this trial for some conflicts here that antibiotics could be stopped early, but then reintroduced early later on.

And so you've got this kind of tension that the initial antibiotic period might not give us the full picture. So we decided, rightly or wrongly, to add up all of the antibiotics that the patient was exposed to from recruitment to day 28 as our primary outcome measures. So the information you're going to see is about that summation. So it's going to be a bit longer than the figures that you're normally used to seeing for antibiotic duration for sepsis.

And to do that we had a significant task, which we did largely electronically, to link records in our system to look at any antibiotic use. So this wasn't just about antibiotic use in the ICU. it could be on the ward or the floor of the hospital. And of course patients hopefully are discharged home. So we had to connect with a community based electronic systems to look at any antibiotic prescription that patients had. A lot of hard challenges here.

But really important to understand this. For our study that peer review, we were also asked to consider what would be our most important clinical safety monitoring. So this wasn't going to be a primary outcome. And I'll talk about that some more shortly. But we decided that we would, having had conversations with patients, that what they didn't want to see was antibiotic shortened, but running the risk of worsening mortality and 28 day mortality was the one that we chose.

Although we have reported mortality out to 90 days as well or cause. So that's our, if you like, chief clinical safety monitoring. But it's not a primary outcome in terms of the sample size. So lots of secondary outcome measures that you can read about to understand mechanisms of clinical effectiveness and safety. We linked our records to patient all cause mortality at 90 days.

And we also measured health care system benefits because certainly in the UK there are significant questions about cost effectiveness of changing practice. And for us PCT is not often routinely used in this pathway. For this purpose. They may be measured for various reasons and often not evidence based. So the results of this trial potentially could change practice.

And we needed to understand the cost effectiveness of that. So just briefly we designed a multicenter three armed individual patient intervention concealed randomized controlled trial. I'll stress the intervention concealment because you've got to think about this. As a clinician at the bedside. They were given daily advice on every patient, but they didn't know the source of that advice, whether a biomarker had been measured or not and which biomarker.

So that's critically important. So in some ways we were sort of slightly disadvantaging the biomarker interventions and probably lifting up the standard routine care of these patients in terms of standard stewardship. So quite a risky trial to do. But that was our design. You can see the three groups there. There were standard approaches to computer randomization. And we were sensitive to the fact that this needs to be stratified by center because of different case mix, which might explain outcomes, the severity of sepsis, whether it's shock or not, in terms of explaining outcomes and whether the patient had been exposed to surgery within the last 72 hours or not.

We wanted some balance in the trial over that because we were encouraging sites to recruit surgical patients as well as non-surgical patients, because often in these studies, we really don't know a full case mix of patients and what these markers may have to offer in those patients. So that was an important part of the trial. Yeah. So the sample size we work with patients and the evidence base based to determine what would be the minimum clinically important difference to see in this trial and to try and capture.

And we agreed with patients that a one day reduction could be important. And we wanted a trial to have confidence that if a one day reduction was present, that we would find it any less than that we didn't think was important. And to do that, based on previous literature, we needed to perform a trial, a large trial of 2760 patients with 920 patients in each group to have confidence about that.

So you can see within this, there is no suggestion that we're powering for safety. The safety that we report is observational within that sample size. So on to the results. Very briefly we conducted it wide scale in the UK across all four nations. And in 42 sites. You can see that the study started in January 2008. Garth to a bit of a slow start.

That's the blue patient across lines over what we were expecting at the time. And I have to say that that was largely because the pro calcitonin had not been set up widely across the NHS. It wasn't something that was performed widely and Jonathan. I'll talk more about that, but I'm very grateful to and we'll mention them at the end.

Three companies that work pre competitively and Roche Diagnostics is one of those to support the NHS to set up those assays and get them underway in the NHS. And you can see the result of that was that our recruitment went up rapidly. And then of course you can see that it stopped. And that was March 2020 when we made a decision in the United Kingdom to pivot all of our critical care and acute care research to deliver important evidence for the management of patients with Covid.

And that's for a separate conversation. But we had to set aside this study. But I'm very grateful to the government, who then reinstated this trial at a difficult time after the first wave of Covid. So the back end of 2020 and we got going again, and they had faith in our community to deliver this, and we completed that in 2024 to full sample size.

So did we conduct an experiment? That's a really interesting question. And what we were able to do at the end of the trial was to open the computer black box that I've talked about, to have a look at what information it was delivering into the trial. And you can see on the right hand side, I'll start with the standard care protocol.

So that was advice that was always delivered to those patients as usual care advice. So there was no stop advice. And I'm very grateful to see in these real data that there is no color other than usual care advice. So it wasn't contaminated with any stop advice. So the protocol seemed to work. Now, what you will see from this information is quite interesting that day one is on the left and day 28 is right out on the right hand side.

And on day one everyone had in the standard care group advice delivered, but you can see advice then starts to tail off. And that's because the intervention period had completed. Clinicians were stopping antibiotics even quite early in the standard control group. And you can see how that cascades over time. So that's standard practice in the United Kingdom.

If we then move on to the Tonin and CRP groups, you can see we try to indicate here with different colors yellow stop advice and blue strong stop advice that there are slightly different patterns emerging between PCT and CRP. So in the PCT group you can see very strong advice occurring relatively frequently quite early. So that's the blue.

Whereas in the group it occurs less frequently. The stop advice in yellow is broadly equally distributed both from PCT and CRP and in about the same amount. So that's really interesting. So opportunities to stop antibiotics in the picked protocol. And you can see this clearly more frequent than in the CRP group. So that's the information that's given to clinicians.

So what are the results. What did the clinicians do with that information. So that's the question. How did they adhere to that advice. So this is about to be put in peer review. It's taken us a long time to get this information out of this automated system, linking decision making to stop with what actually happened in terms of the advice.

But we've got some good data out now. And when we consider the suggest stop. So that's the intermediate stop advice. And I've said that that occurs in similar times in PCT and CRP in similar amounts. Then the adherence is around about 40% and there are no differences between those groups. So that's the average adherence are when that information is given and that sort of concords with what we know in the field from previous results.

So you might say well that's not brilliant adherence. But remember clinicians are faced with the patient in front this is not mandated advice. It's about whether it makes sense to contribute to clinical decision making at that moment. So if we go on to the strong stop advice, it's quite interesting that that adherence goes up. So it's important that clinicians were being given a very strong handle on whether to stop or not.

And you can see actually there might be a small systematic difference between the advice delivered by CRP and PCT, which is very strange because the clinicians didn't know the source of the biomarker in this study. And I think we've done some more analysis of this. And what it appears is that PCT strong advice is occurring more frequently and quite early in the patient pathway, at a time when the patient is much sicker and there is a greater likelihood of that advice being turned down.

So I think that probably explains the difference here, and will present data to peer review very soon in support of that. So just briefly, this is the table one, the kind of patients that were recruited. I'm not going to go through this in detail. We'll give you the paper shortly to have a look at. But these are typical patients in the United Kingdom.

Probably 50% are from medical emergencies, particularly of the respiratory tract, and about close to 30% of surgical patients with surgical sepsis, particularly in sepsis, the vast majority of patients met the sepsis criteria. They had Sofa scores for those that are interested around about seven, and that was balance between the groups and that sepsis and septic shock was present in equal amounts.

Balance between the three groups. You can have a look at these data to look at how it generalizes to your own patients and your own practice. So moving on to the results of the primary clinical effectiveness. the total amount of antibiotics from randomization to day 28 was just over ten days in the standard care group. And remember we're looking for a one day difference as a minimally important.

And so the mean difference in bed with standard care. The one day exists in the 95% confidence intervals. And it is significant. And that difference doesn't appear in the CRC group. And you know there's no significant difference there. And remember clinicians didn't know where these where the advice was coming from. Really important. So in terms of safety around about the average kind of mortality at 28 days was 19 to 20%, that we did work out an observation and observed how confident we were about a non inferiority margin and presented the data in this way and non inferiority was met for the precursor toning group has indicated there and we didn't quite achieve it for the CRP. So there's uncertainty about the safety of CRP. Now the other information that we presented in terms of our observational data on safety and mortality was the Kaplan Meier curve, both at 28 days and at 90 days. And you can see from this there are no differences between the three groups. And there's no statistical significant difference in those in those couple of my curves.

So we declared safety in the trial based on that information and other information. Actually, I'm not going to go through the secondary outcomes. You can look at those. You can ask us about them. But they were balance between the three groups. And finally a whole series of pre-planned subgroup analyzes were conducted. And the effects that I've talked about in terms of total antibiotic duration are not modified by any of these subgroups.

I mean, I'd like to say, therefore, you know, we've firmly shown that our results are applicable to all patients recruited into this study. Of course, that will be overstating subgroup analyzes because as you know, they are hypothesis generating. But I think it is confirmatory evidence that the evidence does apply pretty well to all these subgroups which you can read about.

So in conclusion, in a wide range of critical hospitalized adults with sepsis in our hands, there was about a 10% saving in total antibiotic duration using a daily PCT protocol compared with standard care. And we didn't find that in the CRP group. And that's cut and pasted out of the Jama article, which is now in open access with all the supplementary material.

If you haven't seen it, please scan the QR code there and that will take you to the paper. Sheila, I'm going to progress the slide ready for Jonathan. Correct. Thank you. Jonathan. Brilliant. Thank you Paul. And yes, thank you, Sheila, Bastian, Patrick and Ross for inviting me to talk about the lab aspects of the trial. So yes, I was co-op and National lab lead for the trial.

And I'm currently a consultant clinical scientist in biochemistry at Lancashire Teaching Hospitals. So for my part of this seminar, I'm just going to give you a brief introduction to pro calcitonin and the role of Procol Atoning and CRP in the adaptive trial, which actually Paul has covered really well so far. So I'm just going to add a little bit to that, and then I'll concentrate mostly on the hurdles that we had setting up this trial, and then talk about implementation of a new assay within a clinical biochemistry lab, the role of external quality assessment ECA, both within routine use and within trial use, and what I consider to be the importance of lab representation within trial management groups. So to begin, brief sort of outlook on pro calcitonin. And so pro calcitonin is a 14.5 kilos and precursor protein that in normal health is produced pretty much only in the thyroid as a precursor protein to calcitonin. However, in bacterial sepsis it is raised and it is produced by many different cells pretty much systemically and initially by monocytes and by macrophage activated parenchymal cells.

It does have a short induction time, so it starts to rise approximately 2 to 4 hours post activation. So for whatever the infected response is and it peaks usually at around 6 to 12 hours. And this is in contrast to CRP which has a much longer induction time. It probably starts to rise about 6 to 12 hours after infection and peaks around 48 hours.

So CRP is a much slower response than pro calcitonin, which kind of partly fits into how this trial has worked and the results that have been shown. It does have a short in vivo half life of 24 to 36 hours, which makes it a good marker for monitoring reduction in antibiotic use because it will drop off quickly. And the general consensus was to consider antimicrobial cessation.

When pro calcitonin is less than 0.5 nanograms per mil or micrograms per liter, they are the same units, or there is a greater than 80% drop from peak value.

So the role of pro calcitonin and CLP in the adapt sets is trial. As I say, Paul's already covered a lot of this, so I'll just go very quickly over it. Patients in the trial were randomized into three trial arms the standard care where no biomarker was measured, CRP guided care or PCT guided care. And then daily blood tests were taken from every patient within the trial, regardless of which arm they were in.

The clinical and research team, as Paul has mentioned, were blinded to the test type and the lab were blinded to the patient ID, and the first test must be performed within the first 24 hours of receiving antibiotics. That was a really important start point, and then generally within 24 periods after that, and the algorithm based advice was returned to the clinical team electronically.

So this is lifted directly from the paper that Paul's giving you the link to and that you can read. And we used slightly different to the to what I put before, but we use the pro calcitonin of less than 0.25 micrograms per liter, or a less than 25mg/l for strongly supporting stopping antibiotics and a pro calcitonin, followed by greater than 80% from baseline, or in between 0.25 and 0.5 micrograms per liter.

And the CRP fall from 50% from baseline, which supports stopping antibiotics. And if those criteria, neither of those criteria are met, then the advice returned would be to support usual care, and that will be the advice for every single scenario within the control group. Has no blood test is is performed in those patients. So the hurdles are setting up a blinded trial.

So Paul and I and a lot more of the team discussed this greatly as we were putting together the application to the NIH for the funding for this trial as to how we would approach this, because obviously, in routine clinical trials where the intervention is known and the test is known, then the blinding may simply be for the laboratory not to know who the patient is so that they won't enter into their clinical record, or there may be no blinding at all.

However, in this particular trial, we wanted the lab not to know who the patient was. So if there was any discussion with the laboratory about a particular patient who was in the trial, we wouldn't be giving any results or advice that they didn't want because of the blinding. And similarly, we didn't want the clinicians and the clinical teams to know which test, if any, had been performed so that there was no bias regarding the information that was received.

And there was a requirement, obviously, for these results to therefore remain outside the patient clinical records. But it also meant that there was a huge impact on the laboratory, particularly to make use of the test outside of its routine use, and also to be dealing with blood samples in a very different way to how they would be routinely. And given that laboratories receive thousands of samples every day to make this trial work meant that the laboratory staff, not just like myself as a clinical scientist, but the biomedical scientists within the laboratory had to be on board and to take part in this trial.

So a huge thanks do go out to all the labs that took all 42 sites, because it was a big undertaking from the lab point of view, regardless of anything else. Because of that, we did have the additional IT requirements, which Paul mentioned. The black box. Huge thanks. Do go to Warwick for clinical trials unit for developing that because they did think about it very carefully.

I was concerned particularly that if we were to do just click on a box within the IT system to say that it was a patient that didn't have a blood test, that the advice would be received back really quickly by the teams. And so therefore they know which trial arm that patient within. And similarly, because profiles are in at the time the trial was implemented was not a routine test.

It may well be that that would take longer for the lab to set up and perform the test and report the result back then it might do for CRP, which is a much more widely used test, certainly in the UK. And so those results might return more quickly. And again, it might unblinded the clinical team. So Warwick did work in an average time delay so that the results were returned more or less at the same time, regardless of which arm those patients were in.

So the advice that the clinical teams received, they would never be able to work out which trial arm the patient was in. So that was a really clever way of maintaining that blinding. We did obviously have some failures in this. I was responsible for one of them, because I contacted the Warwick team to discuss a patient that we had an issue with.

I was unaware that they were blinded, and so I told them which the patient was in, and so that patient was slightly compromised. Then within the trial, however, we managed to sort that out, but we asked, maintaining that blinding was difficult. However, again, I'm very grateful to all the lab teams across the UK for honoring that blinding and maintaining it as far as humanly possible, particularly when dealing with analytical factors which could affect the return of results.

If there was a particular analyzer system that was down for maintenance or down because it had some control issues, that meant we couldn't report results. And so there was a delay. It might mean that we would be contacted by the clinical team to say, I haven't had the advice yet, what's going on. And so again, trying to maintain the blinding them was challenging but definitely appropriate.

And we managed to do that. And we got through with the 2760, just over 2760 patients in this trial, which is a huge undertaking.

So in terms of implementing pro calcitonin within a laboratory thanks to Covid, and not many people would say thank you to Covid, but thanks to Covid, PCT was introduced in a great many laboratories as it was part of the monitoring for Covid treatment. And so it was a requirement that each lab have available across the UK. So it meant that, as Paul's mentioned earlier, the trial was paused in the United Kingdom at the start of Covid until towards the end of 2020.

But by that point, many more laboratories had its own and available, and so we could expand the trial to more sites more easily than perhaps we might have been able to before. But implementing the test in a routine use outside of trial use the biggest hurdle it's certainly in the UK that we have to face is finance, and I'm sure this is probably true worldwide.

It would require a business case with a demonstration of tangible cost savings and probably some element of demand management to reduce unnecessary or on indicated testing. The main reason for this is that the cost of pro calcitonin is higher than it is for CRP. So for example, in my laboratory in Lancashire, our SLR costs. So the total cost of the test that that we would charge to the to the hospital uses for pro calcitonin is about 30 pounds sterling, whereas for CRP it's just over five.

So there is a huge cost difference and therefore a cost and expense associated with pro calcitonin monitoring. However, Paul will talk through this later, but there is a cost saving to be made through the reduction in antibiotic use and in other reductions that will be seen through patients who are treated through a precursor turnin protocol. So although there is an outlay needed, there is a cost saving to be made through implementing pro calcitonin.

So from a laboratory perspective introducing a new test, it does require some work. It's not something that can be done quickly or overnight. It needs some method verification with testing of the truth or bias of the test, which is where you use values of known results. This can either be patients that you've already analyzed, protocols to turn in in by another method, or by a method in another laboratory that you are replicating in your own lab.

Or you can use Echa samples. And I will come on to Echa shortly. And then also a measure of precision or imprecision, so that if you are testing the same sample multiple times, that you're getting the same result. We do need to be part of continuing continuing quality assessment and assurance. And as I say, I will come onto that shortly.

And then there is accreditation to ISO 15189 international standards that is required for every laboratory, certainly within the UK. And I would expect worldwide as well.

So the role of external quality assessment. So QA quality assurance covers both internal quality control and external quality assessment. So internal quality control is a task that you would be doing multiple times a day to check that the test is producing the result that you would expect it to. And then external quality assessment is usually taking part in an external scheme provided by one of the EQA providers, and there's a few listed there that certainly are available in the UK.

So there's recast UK next week, and Arabia, which previously was known as lab quality. They all provide pro calcitonin as a scheme under their EQA schemes, and it is essential for in clinical labs for the provision of precise and accurate analyzes to support optimal patient care. So Echa has an important in monitoring because it's showing you how well you are performing within your laboratory group in terms of for all laboratories that use the same analyzer in the same particular test kit, it will show how well you are performing against them, and it will also show how well you are performing against everyone that performs that test, regardless of the methodology that they use.

So peer to peer assessment, which is the same lab, same analytical equipment use is really important. And for showing that you are within the group that you are working with. And if there are any outliers, it will probably point to the being a problem with your test rather than the test itself. Similarly, the cross lab assessment on different analyzers is really important to show if there is a shift or adrift in a particular method.

We, for example, in the UK and I think this has gone worldwide, that has been some issues with some other tests. And so we get regular correspondence from the EQA providers to say there are potential issues with this test and it should be investigated. And then that goes back to the manufacturers as well. So it's a really, really important part of day to day practice within the laboratory that we have these services and systems within in practice to ensure patient safety and the quality of the service, and that the clinicians can be assured and the patients can be assured of the quality of the service that the laboratory is providing to them that will keep them safe.

So just a little bit from me on the importance of lab representation within clinical trials. So the sepsis trial did necessitate significant laboratory input into the design of the trial, mainly because we did have to undergo this blinding. And so working out how that was going to be feasible and how that was going to be possible and required input from me as a laboratory clinician who understood how the laboratory works, but also had insight into how the test would be required within the clinical trial setting.

Not all trials that have laboratory tests as part of them will have laboratory representation on the trial management group. However, I think it is really important and would advise anyone that is looking to set up or design a clinical trial that requires a clinical test, laboratory tests to be performed that they do have laboratory representation on their trial management group, so that the hurdles that I talked about before, both in terms of implementing a test within the laboratory, but also implementing a test within a trial, can be looked at more closely from the laboratory perspective and worked out just how it's going to be possible to do it, particularly if it's not an automated test and requires more manual intervention. There may be a staffing consideration within that, and also if it is an automated one, how those results are returned. Because if they're returned electronically into patient records, is that appropriate for the trial? And if they're not returned electronically into patient record, how do they get back to the trial without the laboratory being directly involved in that design?

So it's really important to involve your local laboratory representative representatives within a trial management group when designing a clinical trial. We're also really useful for the interpretation of test results. So there's not an expectation that clinicians on the wards will know everything there is to know about a particular test. But that's why people like me, as a clinical scientist in biochemistry, are in post, so that we can be that point of advice and interpretation of test results.

But yes, primarily for advice both on how to interpret results but also how to implement them, how to get new tests off the ground, and what the hurdles are and what the difficulties may be and how they can be overcome, and what we can do to work with you to enhance patient care. So just few acknowledgments from me.

The TMG, the local lab leads, particularly because this was quite a feat to get this trial over the ground within the lab setting and the many lab professionals that were required to make the Adapt sepsis trial work. So for me, thank you very much for listening. And back over to Paul, I think, Jonathan, thank you so much. And, you know, it really was a team science event.

This. So thank you so much. So I'm going to just just in the last few minutes talk about the next steps. Because clearly we've got a piece of evidence now that SAT potentially sits on the shelf of a library in the UK. And we want to be a bit more active than that. So the first thing that we've done is to commission an independent piece of work that we're advising on to look at a systematic, a rapid, systematic review to bring in the Adapt sepsis trial results into the broader open labeled literature that already exists.

I don't think there are going to be any too many surprises about the results of of this systematic review, but I think it's worth doing in terms of persuading the broader community that PCT, in our view, has a role in the management of decisions about stopping antibiotics in sepsis. And it'll be really interesting to see what the clinical effect is and whether it's greater than a concealed trial.

So that's a really interesting piece of work going on. You can scan that for the protocol if you're interested. The next I'm not going to go through this in detail. Please look at this later. But we've done a piece of work in the UK asking services in the NHS, whether the results of the Adapt sepsis trial mean that they will change their practice.

And we've done what's called a sentiment and analysis and a test of change. You can read that it's all about implementation science, and there is a clear willingness to change practice in the UK to adopt precursor toning. But the biggest impediment to change, as Jonathan is mentioned, is money. So the final slide here and this is in advance peer review at the moment in a European journal where we've now reported the entire health economic analysis, which strongly suggests that in terms of the approaches that nice use in the UK, and there's been some criticism about how we set standards to compare medical technologies.

But this particular intervention, the cost to produce an additional quality adjusted life here comes in at about 2,000 pounds sterling, which is well below the threshold. We don't have many cost effective interventions in critical care, but this appears to be one of them, and we have a certainty around just over 73% in that result. And there's other results that you can look at when that finally appears in publication.

So I think we've addressed most of the questions here and leave it to you to whether it's going to change your clinical practice. Huge thanks to lots of people, including all the teams that Jonathan engaged with. They're all acknowledged, with individuals acknowledged actually within the publication that I mentioned previously. And here's a letter of governance that went on in the trial and our friends and colleagues at the Warwick Clinical Trials Unit.

Sheila, back to you. Brilliant. Thank you so much, Paul and Jonathan. That has been really, really insightful. We are almost at time, but we have a couple of questions in the chat. We might take a moment just to look at some of those. We can. The first one. Would there be any benefit in measuring both PCT and CRP?

For some patients, having a combined cutoff level based on the fact that one is elevated earlier than the other? I'll briefly mention that and see what Jonathan thinks. We were asked that at peer review and Ethics, and the reason we decided not to do that one was we we'd have to do that across the trial. And so there were some cost issues in the trial.

And secondly, the governance of the black box computer system. We were anxious that we could get into a bit of a mess about confusion, but I think it's a great point and currently untested. Jonathan. Any thoughts? I agree, I think there are certain circumstances in which testing both PCT and CLP at the same time is useful, and that's certainly one of the conditions that we use now.

My trust in when PCT is used, it's not a routinely available test. It's generally only available to our ICU colleagues for monitoring treatment of sepsis. But there are occasions in when CRP might be raised, but there are no clinical markers of infection, there's no temperature, there's no other markers of sepsis, rather. And so there's an unknown cause of a raised inflammatory marker of CRP.

And so I think in those aspects, doing both at the same time, or certainly one after the other is valid and having a combined cut off. I haven't particularly thought about it. And I think, as Paul says, that's an area for potential research in the future to look at whether there is a potential usage of both these biomarkers combined or whether one has superiority over the other in that kind of context.

So I think that's a great question. I don't know the answer to, but I think it's certainly something to worth look into. Brilliant. Thank you. Maybe one more quick question. Have you also included dose optimization using drug monitoring for antibiotics you used in the study. Yeah. So so another great question. So that is our hopefully an next ambition as a trial.

It wasn't incorporated into this study. It was a very clean question that was commissioned. But we have proposed that question for a future study. We'll see what happens. Brilliant. And maybe one last question. Since PKD assays are not yet fully standardized internationally. How did you handle potential variability between different platforms when interpreting trial results? And in your view, does this pose a significant barrier to implementation of biomarker guided care more broadly?

I think that's for Jonathan. Yeah, I was going to say I'll certainly take the first part, if not all of it. That's a great question. And it was something that was discussed at the time. So because the anti antibodies in the test for pro calcitonin are proprietary to Brahms, which is now part of Thermo Fisher and all pro calcitonin assays that were used within the trial.

Use those antibodies. And so although there is a slight variability between the test the tests that were generally used, the cutoffs or rather the the protocols that we used with an 80% drop, you wouldn't that would make no difference which analytical equipment it was performed upon. The only potential was further strongly support stopping advice, because that's based on a pro catatonia of less than 0.25, then we haven't closely looked.

So I haven't closely looked at the data to know whether or not we did see any analytical differences between equipment. However, it will be something that is interesting to look at, but we didn't feel at the time of setting this trial up that it would be significant in altering any results that we might see. I don't think it does pose a significant barrier to implementing biomarker guided such as care more broadly, however, in the absence of an international standard, you are right.

That is something to consider when you are looking at implementing it. So I mean, the only thing I would add, which is a methodological issue, and that is that we stratified randomization by center. So each of the analytical platforms would have had an equal distribution of the three groups. So comparisons were made with a center methodological adjustment. So so so that doesn't at all address the question.

But at least within the trial we try to embrace the possibility of various sources of heterogeneity. And that could have been one of them. Thanks. Excellent. We do have one additional question, which I think maybe as a result of a previous answer, I am curious to know how high resolution mass spec may help in analysis of PCT and CRP, but I think it's that it would help in the therapeutic drug monitoring.

I mean. Yeah, at the moment the tests that are available are immunoassays. The biggest issue with high res mass spec is the length of time it takes to perform the test. They're not quick tests to perform. And so in the context of needing to return results rapidly for a in this sort of context, I don't think that necessarily high resolution mass spec would help, but it is certainly something to look at for the future.

I know, obviously, and Roche being one in particular, have mainline mass spec analyzers in production. And so there is the opportunity in the future to develop assays for these kind of tests that may help, particularly in determining if there are any crossovers and cross-reactivity for these tests within the immunoassay that is currently in use. However, I think the immunoassay in the context of the trial and in the context of sepsis monitoring is is completely sufficient.

Fantastic. Thank you both so much for such an insightful session. It has been absolutely fantastic. I think if anybody has any questions that can certainly reach out if they want to follow up with with either Professor Dark or with Jonathan, please do get in touch and we might be able to put you in contact and have a read of the paper, I suppose, as the key message at this point.

And we keenly await the follow up papers as well. Just note as I say, this is first of a three part series, and maybe the next session will also help answer some of those questions. So it's on purpose. Drug monitoring and mass the trauma tree. What. And that's in December. And then after that in January we have pathogen IP and beated panels.

So if you can join us for the next ten series. And thank you everybody for for joining us today.