SCCM Resource Library-Beyond Cultures: Biomarkers and Rapid Molecular Tests to Improve Antimicrobial Use

Video Transcription

Good afternoon, and thank you so much for the opportunity to speak to you all today.  My name is Dr. Ryan Maves. I'm a professor of medicine and anesthesiology in the sections on critical care medicine and infectious diseases at the Wake Forest School of Medicine.  I'll be talking to you today about rapid molecular tests as a way to improve antimicrobial use in the intensive care unit.  These are my disclaimers. I have no financial conflicts of interest that directly refer to the contents of our presentation today, however.  So the scope of the problem we have is fundamentally one of increasing antimicrobial drug resistance. If there was no antibiotic resistance, none of this would be a problem.  As we all know, delayed active therapy, that is, delayed administration of an antibiotic that treats the infection a patient actually has, clearly increases mortality in sepsis and septic shock.  This study by Dr. Kumar and colleagues in 2009 illustrates this nicely, showing that patients who received inactive therapy had an odds ratio of death up to 10 times greater than that of patients who received active therapy when presenting with severe sepsis and septic shock.  The flip side of this is that increasing antimicrobial use drives increased resistance.  We are aware that there is a certain proportion of patients with severe sepsis and septic shock who have multidrug-resistant infections at the time of presentation.  We engage in an arms race with these pathogens to try to stay ahead of them to make sure that we get patients appropriate empiric therapy before culture data is available.  For example, in this study by Dr. Riege and colleagues, we see that prior carbapenem exposure increases the risk of subsequent carbapenem-resistant infections by up to 7.78 times.  There has to be a way we can balance these two competing needs.  We have difficulty distinguishing between bacterial, fungal, and viral sepsis, in addition to occasionally non-infectious causes of severe inflammation.  This study by Dr. Chappelle and colleagues in critical care medicine last year demonstrated that 35% of patients with suspected sepsis who received broad-spectrum antibiotics in the emergency department did not have bacterial infections.  We have a hard time teasing out who these people are at the time, and knowing, particularly in the acutely unstable patient, that withholding antibiotics, if they turn out to have a bacterial infection, can increase immortality, puts all of us into a bind.  This study by Dr. Self and colleagues published in Clinical Infectious Diseases in 2017 illustrates part of the challenges of using a biomarker-based approach, in this case procalcitonin, in the treatment of pneumonia.  Dr. Self and colleagues identified thousands of patients with community-acquired pneumonia, identified who had viral, bacterial, fungal, and other types of infections, and correlated them with their procalcitonin as a baseline.  Unfortunately, although patients with typical causes of bacterial pneumonia, like pneumococci, tended to have higher procalcitonins, there was wide overlaps between groups, making it difficult to utilize procalcitonin for this initial clinical decision-making at the bedside.  Lastly, we have the problem of culture. Culture is the gold standard in many ways, but it's slow, it takes days, and it relies on technology that is 140 years old.  This is a picture of Fanny Hesse, who is a German microbiology technician in the laboratory of Dr. Robert Koch of Koch's Postulates. She is credited with first developing agar-based techniques for the cultivation of microbes.  This is the basis of the technology, with admittedly subsequent improvements, that we continue to use today.  So what is our solution? Our solution needs to be a few different things. It needs to be rapid, we need to get the results quickly.  The data we get needs to be actionable, it has to be something we can actually make decisions based off of. It needs to be sensitive and specific, of course.  It needs to be capable of detecting antimicrobial resistance in an ideal world, appreciating there may be limitations in that compared to phenotypic methods such as culture.  And so that leaves us with a variety of different methods, PCR, antigen, other ways of detecting genetic markers, biomarkers and the like.  And whatever our optimal solution is, is likely to be some integration of multiple methods to get us to something close to truth that we can use at the bedside.  So when we look at types of assays, we're going to focus on assays that evaluate the upper and lower respiratory tract, in addition to assays that focus on the bloodstream, as well as different ways to rapidly identify different resistance genes.  GI panels are relatively common in many hospitals, we won't focus on those very much today.  Neurologic infections are now able to be evaluated through several excellent multiplex assays. We won't be able to discuss those in any great detail today, however.  Biomarkers are an important part of this, they're an adjunct to these specific techniques that I'm going to describe today.  If you're interested more in biomarkers, I would direct you to Dr. Sameer Chaudhry's excellent talk on the subject at this Congress.  So, obviously, infections of the respiratory tract are a major cause of critical illness. There are, thankfully, multiple commercial assays available, Filmaray, Univero, and others.  These are typically multiplex PCR-based studies that provide semi-quantitative detection of multiple pathogens, as well as ways to detect common mechanisms of beta-lactam resistance.  So MEK-A, MEK-C involved in methicillin resistance. Extended spectrum beta-lactamases, such, for example, as CTXM, the most common ESBL in North American hospitals.  Carbapenemases, KPCs are the most common. There are several others that are common in these panels.  What is useful to note is that these panels do not detect non-beta-lactam resistance.  They don't tell you anything about quinolone resistance or aminoglycoside resistance or anything like that.  Although there are viruses present on the panels for the lower respiratory tract. For example, here on the right, this is the panel for the BioFire Filmaray pneumonia panel.  There are viruses on it. It does not include SARS-CoV-2. When it says coronavirus, that's referring to the pre-COVID seasonal coronaviruses that are detected there.  So this is what's on the upper respiratory virus panel for, again, this is the BioFire Filmaray RP2.1, which is currently under EUA. This is from the FDA website.  So this detects principally viral pathogens causing upper respiratory tract illness, in addition to SARS-CoV-2 in this case.  There are a number of bacteria such as Bordetella and Mycoplasma that it also detects. So these are good studies.  In COVID times, we often wind up having to get these together, particularly in mechanically ventilated patients. So how useful are they?  So this is the INHALE-WP1 study, which is a multicenter UK trial.  This looked at 652 specimens of patients with suspected hospital-acquired pneumonia or ventilator-associated pneumonia.  What they found was using different assays, the UNIVERO and the Filmaray in this case, very high rates of pathogen detection, 60.4% to 74.2%.  And this is compared with only 44% by routine microbiologists. These are clearly much more sensitive. And again, you're getting these results the same day.  They're usually quite rapid and hopefully something you can act on quickly at the bedside.  Now, what are some features that increase the likelihood of a positive PCR or a positive culture?  So in this study, Dr. Randon Collings, this was published in Open Form Infectious Diseases in 2021,  looked at the number of leukocytes present on the sputum gram stain, or in this case, a BAL gram stain. And then also looked at the proportion of those cells that were neutrophils.  What they found was that the more neutrophils you had, the more likely that both your PCR panel would be positive and that your culture would be positive.  But the PCR panel was more often positive at lower levels of leukocyte counts and lower levels of neutrophil counts.  So this is helpful, and this is particularly helpful in patients who are receiving concomitant antibacterial drugs,  that your odds of having a negative culture might be pretty high, but your odds of having a positive PCR are pretty good.  And that can be helpful, again, in guiding decision-making and antimicrobial stewardship. This is from a study by Dr. Foskey and colleagues in Italy,  evaluating patients with suspected concomitant bacterial pneumonia in the setting of COVID-19.  And what they found is that the semi-quantitative nature of these panels correlates pretty well with that of culture,  that both on the high end, when you have greater than 10 to the 6th colony forming units per mL,  and on the low end, where you have either no pathogens detected or 10 to the 4th copies per mL on PCR, that those correlate pretty nicely.  And that is helpful in giving you an idea of what pathogen is dominant and what is potentially just a colonizer in that particular specimen.  So culture-negative detections are also a feature of these, where we give concomitant antibiotics beforehand.  Occasionally, we also have the problem of normal flora kind of overwhelming a pathogen that's present in the sputum.  And what this study by Dr. Buchanan and colleagues showed was that these isolates can often still be detected by PCR,  even with, quote, normal oral flora present or, more frequently, concomitant antibiotics in the previous 72 hours.  Now how often, again, is this actionable? We've used that term a couple times, actionable data.  So Dr. Buchanan's study, again, looked at whether or not the data obtained could affect management.  What they found was that in 48% of patients, there were things identified on a multiplex PCR panel that could lead to appropriate de-escalation or discontinuation of antibiotics.  On the flip side, however, about 16% of the time, they found that clinicians were not necessarily reviewing the results of these assays  or perhaps concerned about other factors in the patient's care that prevented them from de-escalating antibiotics appropriately, despite PCR results.  And what this means is that really good stewardship, antimicrobial stewardship, both by a formal stewardship program, but also stewardship by us as intensivists at the bedside,  needs to be part of the interpretation of these studies, that the best state in the world doesn't mean anything if we don't act on it,  if we don't remember to act on those results to improve patient care.  Now, how about the impact of these viral panels? Well, you know, there's not a ton of data in terms of how they impact antibiotic use,  and in part it's because viral and bacterial co-infections in pneumonia are extremely common. Influenza is the classic example. It seems to be quite rare with COVID.  Meta-pneumovirus is somewhere in the middle. But what we do see is that there's at least some evidence that antibiotic use  can be reduced with the use of these upper respiratory tract viral panels. In this study, patients presented to the emergency department with acute fever and respiratory symptoms,  the use of these viral panels did reduce antimicrobial use, not by a ton, 0.4 days on average, but still that's progress. That is some degree of forward improvement.  Now, does it matter which panel you use?  Now, does it matter which panel you use? In general, the approved tests, be it from PowerCheck, be it one of the BioFire panels,  be it any of the other commercial assays out there, are all pretty comparable. This was a study directly comparing the PowerCheck panel to the Filmaray RP2 and RP2.1 assays,  and what they found is that for the detection of SARS-CoV-2, influenza A and influenza B, that they had extremely high levels of correlation.  So any approved assay should be worth using, depending on which one you can get for your facility and which one you have. Now, how about bloodstream infections?  This is obviously comparable to respiratory infections and often overlapping in terms of its importance to critical care. So conventional automated blood culture systems,  these are continuously monitored. When they turn positive, they stay positive. There's no way to put the bottle back in the machine and start over  in case there's, for example, a suspected contaminant like a coagulase negative staph that grows before perhaps an underlying gram-negative infection.  There is an impact of prior antimicrobial use. If you've gotten antibiotics before, that's going to reduce the yield of cultures. And at best, it's 48 to 72 hours  until identification and susceptibility results are available for the great majority of blood culture systems. And that's only if that blood culture pops very quickly,  as we would see in high-grade bacteremia. So in terms of the types of assays for rapid identification of bloodstream infections,  there are rapid gene detection systems for blood cultures that have turned positive in the bottle already. There is rapid species identification with multi-TOF-based systems  once you have a positive culture growing on a plate. There is whole blood multiplex PCR, which will take a whole blood specimen and identify pathogens present in the blood,  analogous to some of these multiplex pneumonia panels we discussed. And lastly, there's next-generation sequencing, which is perhaps where the future will take us  as we advance our ability to diagnose bloodstream infections.  So rapid blood culture identification systems, there are a number of commercial systems out there, VeriGene, FilmArray, Genmark, and some others.  These typically use something along the lines of a Luminex bead system to detect key genes present in a blood culture isolate.  Now, unlike PCR, these systems don't have an amplification step. To be more accurate, the amplification step is cultivation of the bacterium in a culture medium.  So that's the amplification step. And once you get to the point where there is a positive gram stain off of a blood culture,  that's usually enough genetic material to make this test work. These, in addition to detecting, say, genus-specific markers like for streptococci, species-specific markers  such as for pneumococci, can also detect common beta-lactam resistance markers similar to the ones we discussed in the pneumonia panel.  These have a very high positive and negative predictive value for, say, Staphylococci. You find Staph aureus, and there isn't a MEKA gene present. It is almost certainly not MRSA.  There's a very high positive predictive value for the presence of ESBLs in carbapenemases. That is, if you find a KPC on a gram-negative blood isolate,  it is probably going to be carbapenem-resistant. If you find CTXM on a gram-negative, it is probably going to have an ESBL phenotype.  But the negative predictive value is more limited. It only really detects, or most of them only really detect, one ESBL. And like I mentioned before,  they don't detect non-enzymatic resistance such as porin mutations. That is a common mode of beta-lactam resistance to, say, cefepime or carbapenems in Pseudomonas aeruginosa.  So MALDI-TOF is essentially a form of mass spectrometry. So this is matrix-assisted laser desorption ionization time-of-flight, MALDI-TOF for short.  So this is a way to rapidly identify positive blood cultures. So once you... This would be not positive blood cultures, positive cultures. Once you have culture growth on a plate,  MALDI-TOF can then take that culture colony, put it through mass spectrometry, get a typical nuclear medicine resonance signature that is distinct for each species of microbe,  and give you an ID within a couple of hours. This does not permit antimicrobial susceptibility testing yet, but this is a topic under active investigation.  It's worth noting there are certain species that can't really be differentiated between each other by MALDI-TOF. E. coli and Shigella, for example, have some phenotypic differences,  but they are fundamentally the same species. So MALDI-TOF can't distinguish between those two.  So I mentioned before that MALDI-TOF, as it stands right now, can't detect antimicrobial resistance. But there are studies trying to determine  whether MALDI-TOF could be used as a mechanism for identifying drug resistance. So for example, in this study here, a pathogen was exposed to beta-lactams,  incubated for a period of time, and then the supernatant run through the MALDI-TOF machine. What you found above is, in the presence of a beta-lactamase,  you would see decreased levels of beta-lactam antibiotics and increased amounts of beta-lactam antibiotic hydrolysis products in the lower graph.  Whereas in the upper graph, you see the opposite, where there are plenty of beta-lactams available and very few hydrolytic products present.  So this is an interesting way that MALDI-TOF could be used to further advance our ability to detect drug resistance. So whole blood multiplex PCR,  where we just draw a tube of blood, run it through a couple generations of PCR, generating an amplicon bank, then screening with specific probes,  and then amplifying again for further identification. Is this the holy grail? Is this the thing that will really get us over the top?  So there are a number of emerging commercial systems right now. Early on, there were some monoplex assays for just candida, for example. Right now, multiplex assays,  such as, for example, the MagicPlex sepsis panel as one, are now becoming available. These are rapid, but they are generally less sensitive than blood culture.  So, for example, in this study by Drs. Voromirska and colleagues, when compared against positive blood cultures, this assay was only 29% sensitive,  which is very specific against that blood culture standard. You are less likely to detect something on this assay than you were through traditional blood culture,  potentially because the amount of genetic material from bacteria is relatively low in most cases of bacteremia. Although, interestingly,  it was still more likely to be positive blood cultures in patients on antibiotics, and still could have a role. And, of course, this technology is still very new,  and there are likely to be refinements in the future.  So the real holy grail, if you will, is probably next-generation sequencing. So the idea behind this is that you can measure circulating non-human cell-free DNA from patient plasma.  This is potentially very rapid. It is also, unlike PCR, not pathogen-specific, and would permit you to identify things that wouldn't be present on a standardized panel  with, say, 20 to 30 different pathogens on it. Now, what you do need, though, is a reference or a control bank of control specimens from patients without active infections  to account for the amount of microbial DNA normally detectable in patient plasma. So most people may have, for example, a certain amount of propionibacterium acne  present in their blood in low amounts because we are all colonized with prope in our skin. So in order to screen that out, the patient's serum or plasma, rather,  would need to be compared to this control bank to identify putative pathogens, in the case in this graphic here by Dr. Grumatz and colleagues, Enterobacter,  which is present in the patient but not in the control bank control specimens. So these same authors, again, Dr. Grumatz and colleagues, did an observational series of 50 patients  with primarily abdominal septic shock in Heidelberg and compared them to 20 control patients undergoing elective surgery. And what they found in these patients with septic shock  is that 33% of them had positive blood cultures at baseline, but 72% had positive pathogens identified by next-generation sequencing. And even more useful, 53% of these findings  could have led to changes in therapy. So this is going to be evaluated prospectively in a multicenter study in Germany called the NextGenesis trial,  and hopefully we'll have those results soon and we can really get a look at how to utilize this kind of rapid pathogen detection technology to improve the care of our patients.  Now, culture isn't dead yet. There remains an ongoing need for phenotypic resistance testing, not just because of the complexity of gram-negative resistance, for example,  but also to identify novel mechanisms of resistance as we develop new drugs and as the bacteria fight to keep up with us. There is a limited ability to distinguish  between some pathogens using these current techniques, like I mentioned E. coli and Shigella. Salmonella pathotypes are kind of hard to tell apart through these molecular methods,  but there are epidemiologic reasons that we need culture for contact tracing and the like. All of these systems need close coordination  with stewardship programs, both diagnostic stewardship and antimicrobial stewardship programs to make sure that we act on the data we obtain.  But the potential for rapid identification of these pathogens can improve our time to antimicrobial selection, the accuracy of our selection,  reduce our patients' exposure to drug toxicity, and ultimately, we hope, improve outcomes.  This is my email. Thank you again for the opportunity to speak to you all today, and thank you to the program chairs for the opportunity to participate in Congress.  I hope you all enjoy the rest of Congress. Stay safe. Thanks again.

Video Summary

Dr. Ryan Maves, a professor of medicine and anesthesiology, discusses the use of rapid molecular tests to improve antimicrobial use in the intensive care unit (ICU). The problem of increasing antimicrobial drug resistance is addressed, and the importance of timely administration of appropriate antibiotics is emphasized. Dr. Maves presents several studies highlighting the impact of delayed treatment on patient mortality. He also discusses the challenges in distinguishing between bacterial, fungal, viral, and non-infectious causes of severe inflammation, and the need for a balance between appropriate empiric therapy and antimicrobial resistance. Different methods for rapid testing, such as PCR, antigen detection, and biomarkers, are explored. Dr. Maves focuses on assays that evaluate respiratory tract and bloodstream infections, and the potential benefits of using multiplex PCR-based studies and mass spectrometry for rapid pathogen identification. The limitations and ongoing need for culture-based methods are also discussed. Overall, the aim is to provide actionable and rapid information to improve patient care and outcomes in the ICU.

Asset Subtitle

Infection, Pharmacology, Quality and Patient Safety, 2022

Asset Caption

Increasing rates of antimicrobial drug resistance in the community and in the ICU imperil our ability to care for patients with sepsis and septic shock. In this session, we will review the clinical features that clinicians can use to guide rational antimicrobial drug use, new and emerging diagnostic tests that help us make the right choice of drugs more rapidly, and the data for novel antimicrobial agents, including beta-lactam/beta-lactamase inhibitors and tetracycline derivates, that may support the care of infected patients in the ICU.

Meta Tag

Content Type Presentation

Knowledge Area Infection

Knowledge Area Pharmacology

Knowledge Area Quality and Patient Safety

Knowledge Level Advanced

Membership Level Select

Tag Antibiotics

Tag Pharmacokinetics Pharmacodynamics

Tag Evidence Based Medicine

Year 2022

Keywords

rapid molecular tests

antimicrobial use

intensive care unit

antimicrobial drug resistance

timely administration

appropriate antibiotics

	Society of Critical Care Medicine 500 Midway Drive Mount Prospect, IL 60056 USA Phone: +1 847 827-6888 Fax: +1 847 439-7226 Email: support@sccm.org	Contact Us About SCCM Newsroom Advertising & Sponsorship DONATE	MySCCM LearnICU Patients & Families Surviving Sepsis Campaign Critical Care Societies Collaborative	GET OUR NEWSLETTER
© Society of Critical Care Medicine. All rights reserved. \| Privacy Statement \| Terms & Conditions The Society of Critical Care Medicine, SCCM, and Critical Care Congress are registered trademarks of the Society of Critical Care Medicine.