How can we improve PJI outcomes? And concentrating on improving Amikacin dosing

This month we discuss Amikacin target attainment with Dr Michael Osthoff, and PJI outcomes with Dr Rami Sommerstein

The Amikacin article is here https://www.journalofinfection.com/article/S0163-4453(26)00035-6/fulltext

The PJI article from JOI an be read here https://www.journalofinfection.com/article/S0163-4453(26)00014-9/fulltext

Patient Reported Outcome measures, one study that explored this https://pubmed.ncbi.nlm.nih.gov/38723046/

ASAP study looking at vancomycin for prophylaxis in arthroplasty https://www.nejm.org/doi/full/10.1056/NEJMoa2301401

Thanks for listening to MIC! Feedback is appreciated, email us at journalpodcast@britishinfection.org

Our logo was kindly created by Gabriella Petruso; check out her website at www.gabriellapetruso.com

Transcript

Amy 0:02

Welcome to MIC. More Infection Chat, the podcast for the BIA highlighting articles from the Journal of Infection and Clinical Infection in practice with me, Dr. Amy Bellfield.

Michael 0:14

And me, Dr. Michael Blank, and we're both infection trainees in the uk.

0:19

Today we're exploring why patients have an increased long-term mortality after a prosthetic joint infection, But first, let's talk about improving amikacin dosing in patients with hematological problems. discussing the paper in the Journal of Infection entitled Population Pharmacokinetics of Amikacin in Patients With a Hematological Disorder: A Prospective Observational Study And with us today is Dr. Michael Osthoff, one of the authors on the study. Welcome to the podcast

Michael Osthoff 0:49

Michael. I'm happy to be here.

Michael 0:51

If you could tell us a little bit more about what, what the current landscape is about, uh, amikacin and PKPD or, or pharmacokinetics, pharmacodynamics.

Michael Osthoff 1:00

Sure. I'd probably start with, pharmacology. We've been using Amikacin and many other antibiotics for more than 50 years, and we have been using the same dose, in, every, patient irrespective if he or she is a woman or a man if she's young or old over or underweight. We only. Usually adjusted to the renal function and sometimes body weight. But this is very indiscriminate. and so my interest is in trying to use personalized dosing or to establish personalized dosing for antibiotics. And this is all about pharmacokinetics and pharmacodynamics. So trying to integrate factors that influence. Pharmacokinetics such as renal function, volume of distribution. Albumin concentration for example, and then pharmacodynamics, which is a little bit more difficult because this is what the drug essentially does to the pathogen. And in infectious diseases, this is usually about the MIC, so the minimal inhibitory concentration, which itself is not a very precise measure. Because measurement can be different. Dilution can be different if two people do it on the same day. And then this is a controversial because it's a rich broth medium. It's not really what happens in the body. But essentially I'm interested in trying to use our old drugs more appropriately in the future because they're still needed. Amikacin, maybe. Why did we choose Amikacin? Amikacin is used quite, quite a lot at our tertiary care center for hematological patients. It's a good antibiotic. It. Kills gram-negative bacteria very effectively. But it has a very narrow therapeutic spectrum. So meaning that there is also a risk of toxicity and that's why it's important to dose or to use this drug appropriately.

Michael 3:01

And just about how, dose the drug. I mean, I know there are a number of different possible targets that people use um, PKPD for Amikacin. And you, you do talk about this in your paper actually. Uh, tell us a bit more about that?

Michael Osthoff 3:14

yes sure. So this is again a controversial topic. Amikacin is usually dosed once daily because it shows concentration dependent killing in excerpts are opposed to antibiotic effect. And in former times, the characteristics that most took closely. Correlated with success or killing of the pathogen is the Cmax or the maximal concentration over the MIC. And that has been used in many studies. But we also know that the AOCs for the area under the curve. Might more accurately reflect the total drug exposure over time, and this may better correlate actually with bacterial eradication and clinical outcomes. But this hasn't been used so much in the past because it's more difficult to determine, or in the past at least, because you have to measure the concentration at least twice. Whereas for the maximum concentration, you only have to measure the concentration once. And so, the AO a UC has not been investigated and and used much beforehand, but now with with, for example, patient dosing, I think we don't have to go into detail with that. We can use a limited sampling strategy. And so only measuring amikacin concentration once and still are able to predict the a UC. But in my opinion, the jury is still out. What should be used in the clinics? I think the main message for me is you have to use either one or this is better than only just using a dosage without using a therapeutic drug monitoring.

Michael 4:53

Thank you for explaining that so clearly. And, other factor we, we just before we uh, about um, specific clinical context is um, inhibitory concentration or, MIC, which you mentioned. in your study you rely on u cast's break points for, I should say. Um, Amikacin. Could, you a little bit more about these mean inhibitory concentrations?

Michael Osthoff 5:16

Yeah, usually when you use Amikacin, you use it for empirical treatment. And so, you don't know the minimal inhibitory concentration or the MIC for the pathogen. And even later on only in about maybe a fifth or a fourth, a pathogen is identified. And then an MIC. Can be determined. And so we usually use as a to approach this problem to use the EC cough which is the highest MIC value for an organism without acquired or mutational resistance mechanism, which essentially separates the wild types susceptible population from the non wild type isolate. And for Amikacin the MRC is eight milligrams per liter for enterobacter 16 for pseudomonas aerogen. Which are both very common and important pathogen in hematological patients. But we have to note that this essentially is a worse case scenario because the great majority of isolates, if a much lower MIC which essentially complicates target definition. And in terms of. Amikacin or aminoglycosides in particular? Or in general? There is also the problem that CLSI, so that's the US u cost, so to say has just lowered the aminoglycoside. Break points in 2023. So, we have some discrepancies between the break points in Europe and in United States. For Tactal for example, it's less than four, whereas you cast, it's less than eight. And so, that's something that is has to be taken into consideration. Essentially, we are always. a worst case scenario, which may not always be necessary. For example, in our population, Switzerland, which is a low resistant resistance country we usually have mics of two, are fourfold lower, and this has obviously impact on the amikacin concentration that you might, may want to target. On the other hand, as I mentioned before, the determination is not perfect. It's not a very good measurement.

Michael 7:27

and of course that will have implications for toxicity because if you need to use high doses to get to the. Higher mics then you might be more likely to cause Amikacin toxicity. Is that right?

Michael Osthoff 7:40

That's exactly the case. And so you, you have to somehow de decide what you value more. And this probably every center has to decide depending on their epidemiological, data in terms of the resistance data or even MIC data, whether you really want to go with an MIC of eight or if you're happy also with an MIC of two and maybe one dilution as a safety margin.

8:09

Thank you for explaining that. Now let's, let's move on to the specific population. What, what was it about the population of people with hematological problems that made you want to study them??

Michael Osthoff 8:21

Yeah. So, two reasons. One this is a population that studies have not looked into that population so often. And it is the population in Bosman, in Switzerland where I work, where we use Amikacin most commonly and where we use it. As a standard for febrile neutropenia, for example. And we use it there with a one size fits all approach. And so I was interested if this approach in a population that is quite let's say fragile or vulnerable. Is really sensible. And so for example these patients lack an immune system or have at least a severely impaired immune system. They're often severely ill often with a lot of comorbidities often with resistant organisms, colonized. And so they are sensitive to organ toxicity, organ damage and treatment associated organ damage. And so we wanted to essentially look at our own data in our center to see if we could those Amikacin.

Michael 9:23

And presumably, let's say a typical patient with a ML might get febrile neutropenia multiple times in a year, might receive multiple doses of Amikacin, so it might not be a one-off exposure presumably.

Michael Osthoff 9:35

That's true. So usually in every chemotherapy cycle, they have fe around neutropenia that's quite common. And and I think many centers do not use aminoglycosides in Switzerland at least because they employ an escalating strategy. So they start with a broad spectrum, beta-lactam antibiotic and only add. And AOC glycoside if the patient is still febrile at day three or if a resistant organism is isolated. Whereas we are in Basel we use a deescalating strategy, so the majority of our patients will receive an aminoglycoside. And as you mentioned not only once, but several times during the treatment cycles.

Michael 10:14

The strategy of deescalating, is that because you have high rates of pseudomonas or, or why is that?

Michael Osthoff 10:21

no. As I mentioned that we have a, we are a load resistance country and in base we we have about about 20% of our gram-negative bacteremias are caused by enterobacter and only rarely pseudomonas. I think the main reason why an a deescalating strategy was employed bad experience. So we had I dunno, 15 years ago, we had two or three patients that came in. quite sick. We're treated only with a broad spectrum beta-lactam and happened to have an e sbl l producing organism and essentially died or at least several months in the hospital. So it's all about patient experience in, in our center.

Michael 11:06

Yeah.

Michael Osthoff 11:07

it's very difficult to change this approach because will always say we have a good experience now. We had a bad experience previously. So that is the main reason. It's not about resistance or pseudomonas incidents.

Michael 11:23

I could say in the UK in the centers I've worked with this co cohort uh, myg mono glycosides are commonly used because of antimicrobial resistance. Um, very common practice in the uk.

Michael Osthoff 11:36

Yeah. In Switzerland we have about 12% of all isolates are ESBL producing e coli, ksal, and pneumonia. And this is the case also in, in Basel. But if you break it down, and we've looked at this in several cohorts as well, the chance. Of being infected with an ESBL producing pathogen is quite low. And we do rectal swabs routinely in our hematological patients. So we usually know which patient is colonized by E sbl l producing pathogen. And so for these patients, obviously we use an carbapenem upfront and not even a Minot glycosides.

Michael 12:14

Okay, so let's talk about your study. Now. We've, we've done the background. What, what did you do in your study?

Michael Osthoff 12:22

So this is essentially the second part of an in-depth evaluation of our Amikacin dosing strategy. In a part one, this was a retrospective analysis. We only looked on the impact of Amikacin combination treatment on acute kidney injury. So the toxic aspect. And in the current study we wanted to investigate mainly the pharmacodynamic a aspect. So, so essentially if we are. Reaching our targets sorry target attainment of Amikacin in patients with a hematological disorders. And also describe variables that influence the variability of pharmacokinetics. and so we we included. Hematological patients at risk of developing febrile neutropenia during their course of treatment. So we essentially consented these patients during their chemotherapy. So essentially before they developed febrile neutropenia. And then we sampled them for Amikacin concentration at two time points, essentially. 30 minutes to one hour after the infusion, and again, eight hours later. And then we had a collaboration with a very good pharmacologist vena, gutta from the Children's Hospital who developed a model for our population. And then we essentially had different pharmacological targets that we assessed. So the primary outcome was the c max. Target but we also looked at other targets incorporating the MIC, incorporating the measured MIC for patients where pathogen was isolated. And also investigating a UC targets.

Michael 14:05

Great, and there is a little bit more detail about. The actual modeling and um, you can read in the paper if, want to.

Michael Osthoff 14:11

Yeah, and I have to admit that I'm not a, I'm a pharmacologist. And so we I was really glad that Veta undertook this this modeling.

Michael 14:22

And, what were your key findings?

Michael Osthoff 14:25

Essentially our key findings were that our primary outcome, so a target attainment of a Cmax of more than, 60 milligrams per liter at all time points measured, was attained only in a man, minority of patients. So essentially every fifth. patient which is not surprising because we use rather low dosages and a one size fits all approach. So they, it should be a weight based dosing. So 15 milligrams per kilogram body weight but often. Patients just get a gram of Amikacin. And when we then looked at the body weight dosing, this is essentially even less than 50 milligrams per kilogram. So essentially first finding was the target attainment using a standard target is rather low. But on the other hand, when we used the the measured MIC, so we had 10 patients where we had a pathogen identified. And when we measured the MIC was very low. It was essentially two. And so if we. use this MIC. So Cmax over the MIC of eight most patients would've achieved the target. And so this is really all about what is your target? Is it for example, 60 or 64 milligram per liter? Or is it Cmax over MIC? And if you use the exact MIC or an EC cough. And similarly we, when we use the a OC target most patients exceeded essentially the A OC that that should be targeted.

Michael 16:01

Were you surprised by any of the findings?

Michael Osthoff 16:03

Not really, because as I mentioned we we use standard dosing of Amikacin for all patients. We are not very good in incorporating even even body weight. And we use a rather low dose, so we know from critical Ill. Patient studies that the dosages above 50 milligrams are required to really to really succeed in attaining an ambitious target of for example, Cmax of more than 60 or Cmax over MIC of at least eight. But this obviously comes with the downside and this is then real. Or potential renal toxicity with it, which is not maybe not a big issue. If you only administer Amikacin two or three times, what we usually do, so we don't use Amikacin for prolonged periods of treatment, but it's essentially an empiric treatment until it's clear if an EPL or another resistant organism is isolated or not.

Michael 17:04

And did you find toxicity in your study?

Michael Osthoff 17:06

Not not really. So, essentially all patient achieved a very low concentration at the end of the dosing intervals of less than four milligrams per liter and even less than 2.5 milligrams. Per liter except one, one patient. So, are very low to develop renal toxicity. If the if the concentration at the end of the dosing interval is so low.

Michael 17:33

And to talk a little bit more about how you recruited people to do the study. of the things you mentioned was that, you excluded patients who'd already received Amikacin uh, to the study. Is that right? And did that create some problems with the recruitment?

Michael Osthoff 17:48

So we excluded them if amikacin was already started before informed consent Because then we were not able to measure. pharmacology after the first dose, but only after the third, or second or third. But as I said we usually included patients before they develop febrile neutropenia. So during chemotherapy this maybe creates a little bit of a bias because patients who. Come into the, or came into the hospital directly into the emergency department with febrile neutropenia. Were not included. So these are essentially patients who who come to the hospital to have a chemotherapy and then develop febrile neutropenia during treatment in the hospital. But these were the patients that where we really could study pharmacology after the first dose already in the emergency department, would've been much more difficult for example, during the night, but even during the day because it's essential for these patients to, to have to receive timely antibiotic treatment. And this creates obviously a difficulty for study purposes.

Michael 18:59

And were there any other limitations that are worth noting?

Michael Osthoff 19:03

Limitations in terms of

Michael 19:05

design or

Michael Osthoff 19:06

So first it was a rather small sample size. And this is essentially limiting the analysis of parameters predicting target attainment or clinical outcome. We were not able to study clinical outcomes. For example, target attain is target attainment associated with

Michael 19:23

I.

Michael Osthoff 19:23

or with clinical success. Then these were still patients with quite a good or preserved renal function. so these patients probably still had a low risk of renal toxicity. Baseline EGFR was around a hundred milliliter per minute, which is quite high. So the, our results may not be transferable to patients with function. another limitation, which is not really a limitation because this is just our daily business. A pathogen was identified only in about as I said, 10. 10 patients, or in the minority of patients. And so pathogen specific MIC could only be calculated or could only be used in a minority of patients.

Michael 20:10

For people listening. I was gonna say that's probably a, common experience, right? That people just don't identify the pathogen.

Michael Osthoff 20:18

And probably lastly we only focused on, on, on renal toxicity, but obviously a Minot glyco cytes also carry the risk for oto and vestibular which are both diff more difficult to assess and not dose dependent.

Michael 20:37

We've touched on this a little bit, but do, to what extent do you think this is generalizable to other settings, even within your own hospital? Would you be happy to extrapolate this to other patient groups?

Michael Osthoff 20:51

In our hospital we, we mainly use a OC glycosate really for hematological or Patients. And, as I said, most of our patients were not critically ill. So, and have had preserved organ function. So, translation or a generalizability to other setting is limited. However, as we developed a a pharmacological model, we also were able to simulate, for example, higher dosages that should be used, for example, in critical ill patients. And this is something that, we only touch briefly. On in, in the supplement and in the discussion. But if, for example, if we use a a higher dosage 30 milligrams per kilogram then the consequences are twofold. The target attainment rate increases dramatically. But also the risk of renal toxicity. So the the, the concentration at the end of the dosing interval is much higher. But still these models then allow you to simulate other dosages and pick a dosage that that may be appropriate in your setting. These patients though the hematological patients they have an increased volume of distribution, but not as high, for example, as critically ill patients. Augmented renal clearance is another topic. As I said, our patient population had a quite high EGF also preserved renal function. And so in, in this regard, there are probably similar to, to critically ill patients with augmented renal clearance.

Michael 22:26

And has this gone on to change your practice or, or not?

Michael Osthoff 22:30

we're in discussion with our hematologist as as I said, they are quite how do you say that? Hesitant to change p practice because of experiences in the past. But I think with our data which clearly shows that we are probably not not reaching the targets. We are in discussions to either increase the dosage at least to, to 20 milligrams per kilogram per body weight. Or, and this is something that I, to use an escalating approach. So to stop using Amikacin, at least in, in low risk hematological patients. So for example, in the autologous setting and all oncology patients because they obviously have a lower risk for a prolonged febrile neutropenia episode. So this would be, is my approach. But we are still in, in the discussion.

Michael 23:25

And are there other research questions that this has brought up for you that we, we haven't answered yet? Are there other things you want to study?

Michael Osthoff 23:34

So what we are interested in is to develop models that essentially allow you to define or determine the dose most appropriately. So what we do now is we measure concentration after the first dose, but ideally you should be able to select, the first dose already the first appropriate dose already, and this is all about model informed, precision dosing. So using our. Data that was generated during this study to develop a model that allows you with some inputs, for example, renal function, albumin sex, age, body weight, obviously, to to select the most appropriate initial. first dosage. And then do therapeutic drug monitoring, determine the concentration and adjust maybe only once after the first dose. this is something that we are currently looking into.

Michael 24:34

Interesting. And it does remain a big question about the extent to which outcome data is, is um, reflected after PKPD studies.

Michael Osthoff 24:44

Yes. So this is certainly a topic not only for Amikacin, but for many PKPD antibiotic studies. They're usually retrospective with the limited data set. And so this would be very important because it hopefully answers the question which target is essential. To achieve. For example, what MIC should we use is a c cm max of 60 required or only 40 do should we use an A OC based dosing. And for that, I think larger cohorts ideally prospective cohorts are. Are really required because you will not be able to answer this question with 57 patients as we we did. So ideally we should try to establish large collaborations across countries and try to answer this question which dosing is appropriate, which dosing has an impact on outcome.

25:43

Wonderful. Thank you for explaining this complex topic so well, and, we look forward to some more research coming out of this

Michael Osthoff 25:52

Hopefully I thank you very much for the opportunity.

25:59

And now we turn to mortality after prosthetic joint infections

Michael 26:03

I'm delighted to have Professor Rami Stein on the podcast today, and we'll be talking about, a recently published article in the Journal of Infection entitled, increased Long-Term Mortality of Patients with Prosthetic Joint Infection After Primary Total Hip Arthroplasty, A Large Observational Cohort Study. So I'm delighted to have you on the podcast today. Welcome.

rami sommerstein 26:28

Hello. Thank you for.

Michael 26:32

Let's start by talking about what the background to your study and, what makes you interested in this topic?

rami sommerstein 26:39

So we're having two national programs in Switzerland. So there's the series registry that records all the hip and knee, arthroplasties on the one side. On the other side, we have a Swiss nozzle, the National Center for Infection, control and Prevention. That records, infections for prosthetic joint, surgery. And so there was really a great occasion to merge, these two databases into one, especially because the series registry is linked to the mortality register. In Switzerland. So we have mortality data, and from the Swiss Nossal database there is outstanding data on infection diagnosis. So we're really with many details including pathogens. So we have two very mature and granular databases that were, really optimal to, to be merged, to answer the questions that we were, uh, asking to do. And maybe one more thing, we are really. At the end in preventing, surgical site infections. And what we can see is that within the prosthetic joint infections, we have a really high rate of staph oral infection. And we think these are really, this, uh, staph orus colonization is the main risk factor for a staph oral infection. And this is a modifiable risk factors. So one of the main question was, was staph Oreo infection, was it specifically linked to a increased mortality hazard? And if we could prove this, this could really. Gain some more weight when we say we have to decolonize the patients before they get their implant surgeries.

Michael 28:21

And in your personal practice, would you, decolonize everyone who has staphylococcus colonization prior to surgery?

rami sommerstein 28:31

Yes, we know about the, the staff that what matters is the colonization load. So the more staff you have on your body and the more sites are involved, the more likely it is you get a staph reus, surgical site infection. And from my personal experience is, it is. Not feasible or, or not well feasible to decolonize patients, just if they have staff Oreos, uh, colonization. So in Switzerland we aimed for the option to do universal decolonization for everyone who gets, his knee or hip, uh, ssis. This is the current recommendation.

Michael 29:11

I. Wonder if we could briefly talk about what was published in the literature before your study. What do we know about this topic? Before you published?

rami sommerstein 29:20

So there is some literature on the topic from Danish, from Swedish and from a US cohort, but there, there were methodologically. I think we really, the cleanest cohort because we focused on primary hip. Arthroplasty only, so no fracture related, uh, arthroplasty and no re revisions. So I think we really had a very clean population of elective arthroplasty. And I think this what makes our study different from the other studies. Where, where this was not the case. And, and, and therefore there were also some mixed results. Uh, we saw from the other studies that when you include fracture related arthroplasty, that mortality is higher and the infection risk is higher. And this is something we were able to exclude in our study.

Michael 30:14

I guess the broader, you make your tent, the harder it is to interpret the results.

rami sommerstein 30:19

Exactly. I, I completely agree and I think this was really one of the main advantage that we had, a really high number of observations in our cohort and a really, really clean cohort. So I think this is, one of the main strengths.

Michael 30:36

So let's talk about what you actually did. What did your study involve and what questions did you ask?

rami sommerstein 30:43

So our question was really kind of straightforward. Um, is a prosthetic joint infection in primary hip arthroplasty associated with increased long-term? Mortality rate. This was the main question. The second question was kind of the internal control is, uh, surgery urg due to aseptic revision associated with increased mortality. And then the third question where I was kind of pointing at in the beginning, are selected pathogen associated with a differential hazard mortality risk.

Michael 31:21

And just to clarify, when you say aseptic revision, you mean a revision done for a reason other than infection, so loosening or something other than infection?

rami sommerstein 31:30

Exactly, and once again here, because of our merge bec between the Swiss No and the series platform, we have really valuable data. That makes us quite sure that it's really not an infection. Just as an example, in the, in the cases we identified as infection, 96% had a detection of the pathogen. So we're really quite, um, confident that we can distinguish between aseptic revision and prosthetic joint infections.

Michael 32:02

It, it is amazing that this question has never really been asked before. It, it feels like such a simple question in a way. Uh, it's remarkable that no one has been able to answer this question properly.

rami sommerstein 32:17

Yes, because it depends on the data and we just had the possibility to, to use this data. I mean, this was a concern that, the revision that causes the mortality. So maybe I can already take this a little bit ahead. I mean, in our results now, jumping to the results, a septic revision was not associated with a long-term hazard. And,, I think this really underlines or underscore what a clean, um, cohort we have created here.

Michael 32:49

And what about septic revisions?

rami sommerstein 32:51

So, yes, of course. This is the obvious question. And we saw that septic revisions, they increased the, the 10 year, hazard,, rate for, mortality by a factor. Two. We used the converts, uh, professional hazard model, uh, which involves, time as exponential factors. So this kind of follows the natural, evolution of death with increasing death with higher age. And we really saw this throughout the 10 years that within the model that, that the mortality was, was elevated, uh, throughout the 10 years.

Michael 33:30

What's interesting in your study is you don't say that the revision is necessarily, associated with factors like, age,, BMI, and gender or whatever, but you, likelihood of having that revision is increased if you have one of those factors. Is that right?

rami sommerstein 33:45

Yes, this is correct. So, the cohort was big enough. We had 89,000, uh, plies that we were able to just adjust for, for all the. Comorbidities. And as you mentioned, uh, we did of course stratified analysis within the strata and, and obviously a higher O score, a higher age leads to higher death rate, but the proportional hazard, it remains the same. So I think we have a really. Strong point here to say it's independent of the core variables, but obviously we should keep in mind it remains a cohort. It's not a randomized control study. So, it's not cau at the end, but I think a really strong association where we have trust that, it's a real finding caused by the prosthetic joint infection.

Michael 34:39

Were you surprised by the magnitude of the finding or was this what you were expecting?

rami sommerstein 34:44

I was really surprised in the magnitude when you compare it to aseptic revisions because you could assume that that revision surgery. Could, could hamper the mobility of a patient you have, the, or risk and so on and so on. And, the main surprise was really the difference that aseptic revisions followed. The curve of no revision and the affection. There was a dramatic difference. I think this was. One of the first surprises, and if I can, can go on and we can now talk about pathogen. The next surprise was that our, our primary hypothesis was that, that virulent pathogens such, such as Stef Oreos, streptococci, we have a higher mortality than co bacterium or Coase, negative staphylo oxide. This was not the case. This was actually the second big surprise in our study.

Michael 35:46

I know your study doesn't answer, a question about pathogenesis, but, maybe we could speculate. Why do you think that, QT bacterium that we traditionally think is less virulent, why was it causing the same level of mortality?

rami sommerstein 36:00

This is obviously the, the, the best question you could ask, and I'm afraid I cannot answer it, uh, depending on, on the data of our, of our study. So, so this is really speculative and, and, uh, I mean, one thing could really be that, uh, prosthetic joint infection related. Mobility is, is dramatically decreased. Another factor could be therapy related and there really. Uh, we really want to point out, this is now a hypothesis, so which were the pathogens that were associated with increased mortality? It was ti bacterium. It was staph Oreos, it was the gram negatives. And when you look how these are treated, and we don't have the data, these are frequently treated with Ken alone antibiotics. And the mortality hazard was. Was there for 10 years. It was not just happening within the first year when the infection occurred. It, it was a risk, that stayed for 10 years. So it could be a treatment toxicity related event. But once again, this is, this is speculative and surprisingly str coxal infections, which are frequently treated with penicillins or, or ampicillin, uh, that did not have increased long-term mortality. So. One of the difference could be keen alone treatment. Speculative.

Michael 37:27

the Quin loans and the toxicity from, antibiotics is a good, segue into talking about some of the limitations of your study and what you were not able to ask, uh, or answer, in the study. A big group that you deliberately, miss out is the fracture related infections. Do you think that, uh, could answer the que this question about fracture related infections, using your cohort or would it be very difficult to answer the question?

rami sommerstein 37:54

So unfortunately we cannot because. Because the Swiss no database, uh, does not include fracture related, uh, sty surgery. So they, they only collect, uh, primary sty surgery. So this was kind of excluded by default. It was not a, it was not a voluntary decision. Unfortunately, we cannot answer this question.

Michael 38:17

Okay. Any retrospective data is gonna have some, bias, were there any major selection biases that you noticed

rami sommerstein 38:24

Yeah, I think the main bias versus selection bias, because on the one side, the, the series cohort collects all the autoplasty data and Swiss nozzle. It really depends on the hospital, whether they want to collect the SSI data on TY surgery or on other surgeries. So there is a bias towards. Smaller and specialized clinics to participate in the surveillance, and it could miss out some larger hospitals that have some more complicated primary arthroplasty. I think this is the main bias that we have. So maybe the population that we describe is a bit healthier and a bit younger than the average, uh, D surgery. In Switzerland, but I don't think these are major, uh, is issues major bias.

Michael 39:22

And you also don't stratify by the types of revision. So dare single stage two stage revision. Do you have any sense about, what's, the Swiss revision arthroplasties tend to be? Or, or is this data just not available?

rami sommerstein 39:39

Data was not available in good enough quality, especially that there is frequently not reported in the series registry because they really concentrate on the replacement surgeries. But we know that there is. Relatively consistently chosen for acute infections or acute late infections. So usually if there are symptoms or the distance to the OR is longer than four to six weeks or for, for late infection. If the symptoms are longer than four weeks, then this is really rare that, that their procedure is being chosen in Switzerland. But we don't have the, we don't have this, uh, this data unfortunately.

Michael 40:19

Okay. And there were a small number of people who were excluded from your analysis because of some issues with data linkage. Maybe you could tell us a little bit more about this,

rami sommerstein 40:29

So this data linkage was, being done. Based on probabilistic linkage. So you depend on the hospital, on the intervention date, on the birth date. And even small differences could, could lead to a loss of these patients due to the probabilistic linkage. But I think overall we can say that the date. Of patients that we lost due to the linkage was really, really low. The main issues that one patient was not in one cohort, but it was in the other for, for whatever reason.

Michael 41:04

Right, and, the numbers are very, very low, so the likelihood of it affecting the end result is, is slim, isn't it?

rami sommerstein 41:11

I agree. And, and the reasons are probably at random, so there is no systematic behind it.

Michael 41:16

And, your study looked at the question about mortality. Um, there is a, a big question in lots of PJI studies about, patient related outcome, about quality of life outcomes. Is that data that's collected and, would it be possible to answer that question in the future?

rami sommerstein 41:34

This is a good question. So I'm coming from the Swiss no side where we have the infection and we don't, unfortunately we don't collect the, the proms. I think on the series side there could be some, some data on proms and, I completely agree this would be a excellent addition to the, to the study to. At this data, but we haven't looked at these aspects yet.

Michael 41:58

We touched on the beginning about staphylococcus decolonization. Are there other questions that remain unanswered, that maybe could help to improve outcomes?

rami sommerstein 42:08

I think this is really now the most important question. So we, we have this higher mortality, and I think the obvious thing to do is we should prevent the prosthetic joint infections better and, and, and therefore lead. Live longer in a way. And I think you mentioned the point that we, in the hospital system that I work, we really try to, to do this universal staph Oreo colonization. We have almost no more staph Oreos related, implant infections. So I think this is something that works well. And in a way, each time you see a Staph aureus implant infection, you should really. Question yourself. What happened? Why were we not able to get rid of this modifiable risk factor? I think this is really a key question, knowing that this is associated with increased mortality at the end. We know adding more antibiotics as perioperative prophylaxis. We know it from the TR pill study from Australia, that adding van in, does not change the outcome. So adding more perioperative antibiotics, it's not the thing, so probably really important for this clean surgery is a proper skin disinfection. And then there is one issue, beyond just a skin disinfection. There is probably also some sub cutaneous colonization, by bacteria, by sebaceous glands, by hairs that are, uh, growing to the depth. So we should not forget to do the change of the, of the surgical blade after the skin incision. And there is now increasing amount of evidence that sub subcutaneous disinfection step, uh, is important and could be at least associated with fewer colonization, and maybe also with fewer surgical site infections.

Michael 44:03

So do you, have any plans for any future research in this field? I.

rami sommerstein 44:09

Well. In this field. I mean, now that the, the, the, the studies are merged, I think we are, we're interested in looking, uh, at other aspects. For example, we wanna look if vitamin E substitution in the implant is associated with infections. This is a interesting question. And then a bit further besides our, uh, cohort, I think we really need to learn more about staph Oreos, uh, decolonization. We need to see if, uh, decolonization with Mira and chlorine versus decolonization with disinfectants is interchangeable and, we're lacking on nice randomized, trials in, in this field. And, and I think this is off. Of high importance, high and obviously that the issue of the subcutaneous, uh, disinfection, which has not the same spotlight as the skin disinfection, I think this is an important issue as well.

Michael 45:05

We look forward to seeing some future studies to answer these questions.

rami sommerstein 45:13

So the main learning point infection is associated with increased mortality and. It's not a matter of the, of the pathogen virulence, so it's independent of the pathogen virulence. And the main message is we should do everything possible and reasonable, of course, to prevent, uh, prosthetic joint infections.

Michael 45:36

That's a very important learning point to take away and I really appreciate that you have found the time to come on the podcast to talk about this today. So thank you very much.

rami sommerstein 45:44

Thank you. You're welcome.