SCCM Resource Library-Do You Believe in the Magic of Rapid Diagnostic Tests?

Video Transcription

Thank you to the Critical Care Congress for inviting me to speak today. My topic is the utility of rapid biomarkers, and I've used procalcitonin as my case in  point for three reasons. Number one, it is extensively investigated. Number two, it is very commonly used.  And number three, it still remains pretty controversial with respect to its utility. I have no disclosures.  I serve on the IDSA sepsis task force and CMS sepsis technical expert panel. I'll be discussing sepsis and antibiotic use as it pertains to procalcitonin, significance  of procalcitonin in COVID-19. I'll also try to critically assess the literature on starting and stopping antibiotics in respiratory  tract infections and antibiotic discontinuation in critical illness and sepsis, as well as discuss guideline recommendations and recommendations for the future.  In a point prevalence study in 2007, we learned that globally at any given time, 71% of patients in the ICU are on antibiotics.  Ten years later, this study was repeated and not much has changed. We use a lot of antibiotics.  Despite many advances in sepsis over the years in terms of diagnostics and therapeutics, we still do not have a pathologic or diagnostic gold standard for sepsis.  Now as a result, there's a great degree of uncertainty and that drives excessively broad spectrum antibiotic therapy for patients with sepsis.  In this study that was led by Chanri that we published in JAMA Network Open recently, among patients with community-onset sepsis, there's a lot of broad-spectrum antibiotic  use and yet a relatively low level of resistance. Over the years, there's a number of drugs that have been tried as host-directed therapies  and most of them have failed, and experts suggest that part of it could be the lack of a rapid and specific diagnosis.  It makes no sense to use 21st century technology to develop drugs targeted at specific infections whose diagnosis is delayed by 19th century methods. So are biomarkers the answer?  Biomarkers broadly defined are markers of the biological process or state. Now if you're going to use biomarkers, you have to ask yourself the key utility question.  Is this going to inform me with any more information than I can already gauge by conventional clinical judgment?  And if so, does this information improve patient outcomes and or is cost effective? Now in 1988, the idea of sepsis biomarkers was first introduced by Hoyt and colleagues  who showed that lymphocyte differentiation predicts sepsis. And over the years, there's a number of molecules that have been tried as potential biomarkers  in sepsis and there's really only one that has made it to the guidelines and has been used for guidelines for sepsis and antibiotic stewardship and that's procalcitonin.  Now procalcitonin has been in use for many years in Europe and it was first FDA approved  for risk assessment of sepsis on ICU admission in 2007 and then subsequently for prognostication  in sepsis in 2016 and then to guide starting and stopping antibiotics in lower respiratory tract infections as well as discontinuation of antibiotics in sepsis.  And I'm really going to focus on these two label indications and really critically assess the literature to see what the clinical utility and real world use is.  Procalcitonin is released in a constitutive fashion from all cells in response to bacterial  infection and on the other hand viral infection by virtue of interferon gamma actually inhibits  this response and we think this is the reason why you see an elevation in bacterial infection  but not necessarily in viral infection. Massacott and colleagues first described in infants an elevation of BCT in severe bacterial  infections compared to other viral infections and since then there have been a slew of studies  that have come out and procalcitonin outperforms other biomarkers by virtue of the fact that  it has a better trajectory for real world use because it is detected in two to four hours since the onset of infection and it peaks at about 12 to 24 hours.  As against CRP that has a sort of a slow rise as you can see here. BCT is not the perfect sepsis marker.  As you can see here in the meta-analysis of 30 studies, 3,000 critically ill patients,  the pooled sensitivity was 77%, pooled specificity was 79% and the area under the curve is 0.85. So not perfect but a perfect test does not exist. So it's something.  For those of you who use this test, it's important to know that you can get an elevation in procalcitonin  as a result of trauma, surgery, hypothermia, CKD, neoplasms, as well as some non-bacterial infections such as malaria, dengue, and now severe COVID-19.  Let's understand this relationship. So early on in the pandemic and I think to some extent even now, we have been using a  considerable amount of antibiotics for COVID-19 patients on admission even though the co-infection rate has been shown to be rather low.  In unpublished data here that we queried in Premier Healthcare Database, you can see on the y-axis you have the percentage of COVID-19 diagnosed encounters that received antibiotics  and you have all the hospitals lined up on the x-axis and you can see wide variation in the use of or the extent of use of antibiotics and this really goes to show that we really  don't know whether it's beneficial or not. And wouldn't it be nice to have a biomarker to be able to answer this question?  However, procalcitonin is independently elevated in severe COVID-19 as shown in this meta-analysis  of four studies and the possible theory behind high PCT levels in severe COVID-19 is that  interferon gamma inhibits procalcitonin secretion, but chronic ailments such as diabetes are  associated with a high risk of severe COVID-19 and often also exhibit poor interferon gamma  responses and maybe it's the de-repression that results in high procalcitonin concentrations. Now whether this is a true relationship or indirect, we're not sure yet, but what I can  tell you is that we don't have data yet on the utility of procalcitonin in ruling out bacterial infection in COVID-19 and further research is necessary.  Now research has shown that the distributions of procalcitonin levels in gram-negative sepsis  are higher in general than gram-positive and while this is helpful in an academic sense, I don't think that it helps me very much clinically.  We queried the Cerner Health Facts Database for hospitals that did procalcitonin on admission  and of the 60 hospitals that did that and we identified 42,000 patients with bloodstream infection on admission and 20% met the CDC sepsis criteria and we found that 10% of the  patients who presented with suspected infection or sepsis were PCT tested on admission and the question is, was it ordered to diagnose sepsis?  Was it ordered as a baseline to examine PCT trajectories in the future or were the doctors simply checking off an order set?  This is unclear and unknown and I think it's important to know in the future, but what we could find out was that severity of illness was what was really driving the variation  in PCT rather than organism or site and in general, there has been an interest in seeing whether PCT can help prognosticate patients in critical illness and sepsis and as I've  shown you in 2016, it was FDA approved for this purpose. A prospective study, the MOSES study that looked at emergency room admits with sepsis  examined daily PCT levels for five days and what they found was that the patients who had a decrement in their PCT of more than 80% had a much higher chance of surviving  the hospitalization than those that did not, but then again, I wonder what to do with this information and the authors proposed that maybe a non-decrease in PCT over time may  mean that we're not treating the infection appropriately, there could be resistance or  there may not be infection, but this sort of study has been done. It's the PASS study or Procalcitonin Survival Study.  It's from Denmark and 1,200 critically ill patients were randomized to either receive daily procalcitonin measurements such that if it did not decrease below one over a period  of five days, it resulted in treatment escalation and the comparison was standard of care and  the key finding was that there was no difference in survival at 28 days and instead, we actually  saw signals of harm, greater AKI, more dialysis and mechanical ventilator needs and increased ICU length of stay.  So as such, procalcitonin guided escalation in the ICU management cannot be recommended at this time. How about starting and stopping antibiotics and pneumonia?  So myriad trials have been done on this question over the last few years and this very nice Cochrane meta-analysis nicely summarizes the results and what I want you to focus on is  that the fact that this is an individual patient meta-analysis, which makes us richer, 26 RCTs  were analyzed and they looked at either procalcitonin mediated initiation or discontinuation or both against standard of care and I want you to focus on two signals here.  One was that there was a signal towards improved survival associated with the use of procalcitonin  and also decreased antibiotic use by two days on average and both of these were deemed high quality evidence.  Now while I think I'm convinced that PCT guided antibiotic use probably is not harmful in patients with suspected pneumonia, I think that the author's conclusion was flawed because  they conflated initiation and duration and he said that the use of procalcitonin to guide initiation and duration of antibiotic treatment results in lower risk of mortality, lower  antibiotic consumption and side effects. However, impact of initiation and discontinuation was inextricable because it wasn't exactly clear how the clinicians used the algorithm.  Also survival benefit is largely driven by one study, De Jong, and that study is a discontinuation study.  Also two other studies on procalcitonin guided initiation in critically ill patients, which includes a number of patients with pneumonia, did not show any benefit.  And so in summary, there's insufficient evidence also to recommend PCT to start antibiotics for community acquired pneumonia.  Now there's another issue and that is length of therapy. The Medicare database had data from 2012 to 2013 that was mined by folks at the CDC that  showed that the median length of treatment for pneumonia during that period, during that time was 10 days.  Now we know that there's a growing penchant for shorter courses because of emerging evidence that shorter courses of therapy, including in pneumonia, are not particularly harmful.  And the question is, would these older PCT trials repeated today still minimize antibiotic exposure to the same extent given our current practice?  Two recent RCTs of ED patients who presented with lower risk retract infections, one from the US and one from the Netherlands, both from 2018, did not show any improvement in  survival nor did it show any decreased antibiotic exposure. Notably in the New England Journal study, the antibiotic use was quite low in both groups.  That does suggest that maybe, you know, usual care in terms of duration of antibiotic therapy is changing over time.  That does impact the sort of utility of procalcitonin today compared to in the past. What do the guidelines say on procalcitonin use in pneumonia?  The ATS-IDSA guidelines for community-acquired pneumonia do not recommend PCT for starting antibiotics.  The IDSA-ATS guidelines for HAP and VAP also do not recommend PCT for initiating antibiotic therapy.  Both are strong recommendations of moderate quality and the European guidelines for HAP  and VAP actually do not recommend serial PCTs even to reduce antibiotic duration when the anticipated duration is about 7 to 8 days.  Now how about procalcitonin-guided antibiotic discontinuation in critical illness and sepsis? There's no shortage of data here.  Thirteen meta-analyses were done on PCT-guided discontinuation in infection, critical illness, and sepsis over the years.  And prior to 2017, it was a mixed bag of these three entities and the general message was that procalcitonin-guided discontinuation actually decreased antibiotic exposure without  harming patients, so without any survival benefit. Then in 2017 and 2018, two more meta-analyses were published specifically in patients with  sepsis that also did not show survival benefit and showed low certainty of evidence with high risk of bias around the antibiotic duration question.  However, to our surprise, there was actually two meta-analyses that published that PCT-guided discontinuation in critical illness actually saves lives.  And I think for intensivists in general, if we see a survival benefit, we tend to be moved by that and survival benefit tends to change or modify practices over time.  And so is the survival benefit here really the game changer? And to answer this, we did a 14th meta-analysis.  This was done by Dr. Dominique Pepper and myself, which was published in Chest recently. And here, what did we do?  We were looking at the 16 trials from 2008 to 2018 with around 5,000 patients analyzed  who had critical illness and these were pure discontinuation trials, so PCT-guided discontinuation versus standard of care. And our primary outcome was mortality.  We were trying to see just whether there was a mortality benefit and the secondary outcome was antibiotic duration and length of stay.  Now what is different about this meta-analysis is that we were trying to look at certain special factors that may impact outcome.  And those are, is the signal different in critically ill patients with sepsis? Is the signal different whether clinicians used or did not use the algorithm based on adherence?  Was there a difference in industry-funded and non-funded studies? Was there a difference if PCT was combined with CRP or not?  The first thing we realized when we did our lit search was that all the trials, the PCT results were unblinded and that's just by virtue of design.  If you want to test an algorithm and performance and the clinician needs to see the results.  So while we acknowledge that, there could have been a hawthorn effect and for that reason I think that all these studies were considered high-risk of bias.  Now we did find a survival benefit associated with PCT algorithm as our primary finding. However, there was no survival benefit noted among subgroups of RCTs with patients with  sepsis when the algorithm adherence was more than 80%. So mind you, if the clinicians were using the algorithm more, we did not see the survival  benefit and that really decreases sort of the PCT attributable survival benefit. Also we did not see a survival benefit in non-industry-sponsored studies and when PCT  was used exclusively. Now in the study that drove the mortality benefit, which is a SAPS trial by DeJong,  of 1,500 patients, prospective open-labeled RCT, 15 hospitals in Netherlands and they looked at the intervention of non-binding advice to discontinue antibiotics if the PCT  was greater, if the decrease was greater than 80% or it was less than 0.5 against standard of care, they saw a dramatic survival benefit and a decrease in lower antibiotic duration  by two days as well as decreased costs. Now it's important to take a deeper look at this because it's really driving all the survival benefit.  What we found was that there was no algorithm compliance. So more than 50% of the clinicians were disregarding the procalcitonin result and this is important.  I'm glad they did that because it's important to have clinicians use their judgment as primary,  but what it does is it further decreases the strength of the association or attribution of procalcitonin to the survival benefit.  Also in the modified intention to treat analysis, 29 patients were excluded whereas the fragility  index of 9 shows that there are only 9 patients that are required to move to the other side to change the direction of the results and mortality benefit was an unexpected finding  in this study. The authors theorized that the plausibility for survival was antibiotic adequacy, timely  recognition of alternative diagnosis, lower antibiotic toxicity, and all these may have  lowered mortality, but it still remains speculative and we actually need to do trials to really determine what the reasons were for the survival benefits if such a benefit exists.  Also we found that if there was high adherence, there was not an association with lower antibiotic  duration, which I would say is kind of paradoxical. Cost effectiveness is a weak argument to steer ICU use in PCT as shown in a recent study  where ICU-driven decreases in costs because cost is driven mostly by length of stay was really relatively less compared to the benefit on the ward or in the clinic.  The surviving substance guidelines do say that you could use PCT to assist antibiotic discontinuation in patients who initially appeared septic as well as those who initially  appeared septic but subsequently have no evidence of infection. The pediatric guidelines say there's a knowledge gap.  The NICE guidelines say there's insufficient evidence to recommend routine adoption of PCT in the NHS and the IDSA and SHEA guidelines say that serial PCT measurements to decrease  antibiotic use in ICUs is fair game, but all of these are weak recommendations, but they  do say that one can develop internal guidelines to assist clinicians in interpreting and responding  appropriately to PCT results, a sort of stewardship plus, and in a before-after study in a single  center PCT implementation was associated with an incremental benefit to stewardship practice,  but from our experience in vitamin C I think we've learned our lesson that single center before-after studies are to be taken with a grain of salt and ideally I'd like to see  a cluster randomized trial of PCT versus no PCT hospitals in a, you know, in a system that already uses antibiotic stewardship to see if there's any incremental benefit.  A couple years ago we did a study using premier healthcare database that was the first to look at the real world patterns and use and utility of procalcitonin in patients with  sepsis and we found some marginal decreases in hospital length of stay, antibiotic exposure  and similar mortality risk. However, the take-home point from that study was that both groups received 16 days of antibiotics on average, and so what we may think might be happening  in the real world is different from what actually was happening, and for the most part you can't  really make any inferences from these data, they're outdated, and we need current day  reappraisal of these practices. So in conclusion, PCT has limited utility for antibiotic initiation  in lower respiratory tract infections and sepsis. PCT cannot be recommended for treatment escalation in the ICU. PCT-guided antibiotic discontinuation in lower respiratory tract  infections, ICU, and sepsis does not appear harmful, but it's unclear if there's an impact  in antibiotic discontinuation in the new age with shorter antibiotic durations, and there's  insufficient data to recommend PCT for guiding antibiotic use in COVID-19. Future research is needed on how providers should respond to PCT, how they respond to PCT results. This  is critical to really understand whether there is a utility. Also, we need to understand  mechanisms around any survival benefit attributable to PCT, and we need a reappraisal of the utility  and cost-effectiveness of PCT in an environment with shorter present-day regimen. Thank you. Hi, my name is Simon Lam, and I'm a clinical pharmacist from the Cleveland Clinic. Thank  you for attending this session today, and thank you to SCCM for the opportunity to speak today. I have no conflicts of interest to disclose. So to start, why is antimicrobial  stewardship important? I think it's irrefutable that antimicrobial resistance is going up  internationally and that the use of antibiotics directly contributes to antimicrobial resistance.  This is a global map of countries where carbapenemase are considered widespread. In full circles,  in this case, are bad. In the case that it's endemic to the whole country, open circles are not good. It equals significant regional spread, and stars indicate sporadic outbreaks.  So we see countries like Greece, they're endemic with three different kinds of carbapenemases.  In the U.K., we see significant regional spread of four different kinds of carbapenemases. In the U.S., we see that KPC is considered endemic. So this is another map in the CDC  with states that have reported CRE from 2017. I did a very similar talk a few years ago, and there was at least a few states without CREs at that point. But as of 2017,  every single state in the United States has now reported the presence of CRE.  Excuse me. Aside from resistance, of course, antibiotics by themselves are not 100% safe.  This study is by Tama and colleagues from JAMA. They evaluated 1,500 patients who were admitted to general medicine wards who were given antibiotics. They evaluated them for 90 days  after receiving the antibiotics, and overall observed that 20% of patients experienced at least one antibiotic-associated adverse drug reaction. The most common ones were GI,  renal, and hematologic. And the authors found that, taken all together, for every 10 days of  therapy, the risk of ADE associated with an antibiotic increases by 3%. Importantly, the authors also found that some of the ADEs were observed in patients where  antibiotics were not indicated. So there's many reasons why we shouldn't give antibiotics, but at the same time, as clinicians, we are bombarded with arguments that we need to give  the correct antibiotics as early as possible. This is a study that we're all aware of from Dr.  Kumar and colleagues. They evaluated patients with septic shock. It states that for every hour that we delay antibiotics, mortality worsens by 7.6%. Based on this and other works,  the Surviving Sepsis Campaign does state that, does have a strong recommendation that states we should try to administer IV antibiotics as soon as possible and within one hour.  So as clinicians, we're left with this conundrum, an internal dialogue, where on one side it says  that correct antibiotics saves lives, but on the other side it states that overuse of antibiotics  leads to badness. This internal dialogue is also mirrored in our guidelines and best practices.  The CMS Step 1 criteria states that antibiotics should be administered within three hours of  meeting criteria for severe sepsis, and at the same time, Surviving Sepsis Campaign guidelines recommends for daily assessments for de-escalation of antimicrobial therapy.  So how are we doing in this balancing act?  I would say that currently, the side that says antibiotics saves lives is winning. A cross-sectional study found that 71% of all ICU patients were receiving antibiotics.  According to a study by Lute and colleagues, 30% to 60% of antibiotics prescribed in the ICUs are considered unnecessary, inappropriate, or suboptimal. The CDC also reported a 37%  rise in the use of carbapenem and 32% rise in the use of vancomycin.  So we've set the stage for this tension with giving versus stopping antibiotics. For the  remainder of the talk, I want to explore how rapid diagnostic tests may help in stewardship efforts.  So to start off, what do I mean when I say rapid diagnostic tests? This is a traditional culture  and susceptibility report that I'm sure all of us are familiar with. We get the culture,  the pathogen, the MIC, and also the interpretation of the MIC to determine whether the antibiotic is  sensitive, intermediate, or resistant. There are a lot of advantages to this report. We get the  whole picture, all the antibiotics at once. And this is phenotypic results, which you can likely  rely on in interpreting whether or not our antibiotic would work for the patient. And we  also have MIC information, which we can use to optimize the pharmacokinetics of our antimicrobial  regimen. However, the downside is that the process takes time. So usually with a positive  culture and gram stain, it would take 24 hours for a subculture and then another 24 to 48 hours  for full susceptibility results. In other words, the total time from culture positive can be 48  to 72 hours before we get full susceptibility results. However, there's now a whole closet full  of new microbiological diagnostic tools that could get us closer to the results more quickly. Sometimes from positive cultures to identifications within only a couple of hours.  Importantly, these tests, while very sensitive and very specific, they do not give us the full  antibiotic susceptibility results. So if you were to use any of these tests, you have to commit  yourself to using the information and doing a little bit of deductive work to adjust your antibiotic and understand that there's the possibility that further adjustments are  needed when full susceptibility results are available. There is now also a commercially available path platform that can get us full susceptibility results earlier, within seven  hours. And for the purpose of this presentation, I will also count this as a rapid diagnostic  system. However, there's very little clinical experience in the literature describing this particular accelerated phenotyping system. So these rapid diagnostic tests employ all  different types of technology, but underlying all of it, it's a way to get to the genetic or molecular identity of the pathogen or resistance gene information. And we'll quickly  go through a few of these. The first one is PNAFISH, or peptide nucleic acid fluorescent in situ hybridization. What it does is it links up fluorescent probe to certain species RNA,  and when read under a fluorescent microscope, will identify the species. The example here is we have  a positive gram negative pathogen. We're going to utilize test probes for nucleic acid for  pseudomonas or E. coli, where E. coli will show up as green and pseudomonas will show up as red  under a fluorescent microscope. Currently, this technology can only be used for organism identification. There's no commercially available product that target resistance genes.  Another system is called MALDITOF, or matrix-assisted laser desorption ionization time of  flight. In this method, you plate the pathogen, and then perform a laser-induced spectroscopy,  and then compare the results patterns to known standards. This can be widely applicable to both  gram positive, gram negative fungal pathogens, and also resistance genes, and it's really only  limited by the availability of reference standards. Another method is called microarray,  and is performed using nucleic acid extraction and hybridization, and targets only a specific micro-segments of DNA within the genetic material, and it also has wide application for  gram positive, gram negative, and also resistance genes. So rapid diagnostic tests have been evaluated  extensively, and it's well known that they can shorten time to species and also resistance  identification, and it also has high sensitivity and specificity, upwards of 90 to 100 percent.  You can also decrease time to change in antibiotics, and this can be both in escalation  and de-escalation, and most studies have found that they're either cost-neutral or cost-effective.  Of course, there are some limitations as well. The current systems still rely on culturing,  or gram stain and growth media, and not from whole blood. Most of the applications are only  for blood cultures. They have the inability to detect polymicrobial cultures, and there's an  absence of full susceptibility information, with the exception of the accelerated phenotyping program that I previously mentioned. However, many of the  rapid diagnostic tests can test for genes associated with antimicrobial resistance.  So now that we've discussed broadly what rapid diagnostic tests are, I think potentially these rapid diagnostic tests may have a role during all phases of antimicrobial treatment.  Dr. Kadri already discussed the use of procalcitonin, so I won't belabor the points here,  but I do want to focus on the use of surveillance cultures, which frequently employs one of the  technology that we previously discussed, and also the use of rapid diagnostic tests once clinical  cultures are positive. Of course, no antimicrobial stewardship efforts are sufficient if we're not fully committed to infection control, hand hygiene, and also isolation.  We need to talk about the use of rapid diagnostic for surveillance cultures, and not just its application when clinical cultures are positive, and that's because  majority of the sepsis that we treat in the ICU are culture-negative, and we might be able to  leverage our surveillance culture results to make better treatment decisions. One thing to keep in  mind is that, not surprisingly, administration of antibiotics prior to obtaining cultures lead to decreases in positive cultures. So when you've made your decision to start  empiric antibiotics, or antibiotics have already been started and it's day two to three of therapy,  and the cultures are negative, perhaps surveillance cultures taken earlier in the ICU course may be  helpful. This is a study that evaluated the predictability of MRSA nasal swabs. They evaluated  over 12,000 patients from six ICUs over an 18-month period, of whom only 3.7% of them had a positive MRSA swab. Roughly a third of the patients were treated for an MRSA infection.  Importantly, the incidence of MRSA infection among patients who are treated is really low, only 90 patients or 2% of the patients end up having an MRSA infection.  Given the low prevalence, it is not surprising that having a negative MRSA swab is incredibly predictive of not having an infection with an NPV of roughly 99.4%, whereas a positive  swab actually doesn't tell us much with a PPV of 32.3%. In other words, if you have a negative MRSA swab, you can almost certainly either not  start vancomycin empirically or deescalate from vancomycin. As it turns out, the same principles can be applied to pretty much all surveillance culture information that you have.  In the previous study, you might be thinking that the rates of both colonization and infection are really low, and that is not really representative of your ICU.  Unfortunately, data like the previous study for other organisms are rare in the literature,  but I did find this one study where they implemented surveillance cultures to help combat carbapenem resistance within their trauma ICU.  In this particular hospital and trauma unit, carbapenem-resistant Acinetobacter was endemic, and the trauma ICU had the highest prevalence among any unit in the hospital.  In this study, they performed rectal swabs or respiratory cultures if patients were intubated on ICU admission and also weekly while they're in the ICU.  The prevalence of Crab infection in this unit was much higher, 60 of 364 patients or about 16% of patients eventually had carbapenem-resistant Acinetobacter infection.  Even with these prevalence rates, a negative swab had a negative predictive value of over 90%, and a positive predictive value that was slightly better, but still nothing to  hang your hat on. In other words, even in a setting where organism is endemic and highly prevalent, the negative predictive value of a negative surveillance swab is really high.  And I'm not saying that for someone who's in shock, that you shouldn't start empiric antibiotics if the swab is negative.  However, if someone is not in shock or when the patient is otherwise stable, it's probably an okay strategy to rely on surveillance cultures.  As I mentioned, data on other organisms are not readily available, and this is largely because many of the studies only included patients with actual infections or positive  swab, so a prevalence rate cannot be calculated. However, in this meta-analysis of VRE surveillance swabs, which included 37 studies, they found  that VRE rectal swabs on admission was positive in about 9% of patients. They were once again not able to determine the infection based on colonization since  the studies differed on whether they included only positive swabs or positive infections.  In the review of the studies, the authors did note that the risk of VRE infection among non-colonized patients is negligible or less than 2%.  So if your hospital routinely obtains colonization information, I would urge you to use that information more than just for infection control purposes.  You might be able to use that information to guide your empiric therapy and also de-escalation strategies.  So we've gone through how colonization swabs, which frequently employs rapid diagnostics, can be used in determining the risk for actually having certain infections.  This might once again be helpful for decisions for empiric therapy and also de-escalation. Next we'll go through how rapid diagnostics can be helpful when clinical cultures are  actually positive. As previously mentioned, traditional susceptibility results can take 48 to 72 hours after cultures are positive.  However, there's currently another system on the market that may be able to shorten that time.  This system is fully automated and it combines PNA-FISH technology with automated microscopy.  It can shorten time to full identification and susceptibility results to only seven hours.  This study evaluated this platform's performance and was done in the ED and ICU of Barnes-Jewish  Hospital, where they implemented this accelerated phenotyping method to obtain susceptibility results.  They ran these accelerated samples side-by-side with standard-of-care microbiologic results over a 14-month period, and not surprisingly, they observed that time from blood culture  positive to identification and to full susceptibility results was significantly shorter associated with the accelerated platform.  When looking at the agreement between the accelerated method and traditional susceptibility testing, there's agreement in majority of the tests.  There was no very major mistakes, which was defined as the accelerated method calling something sensitive when the standard-of-care method determined it to be resistant.  There were some major disagreements, which is the opposite, where the accelerated method called something resistant when it is sensitive.  While major disagreements are bad, it is at least a little bit more conservative clinical approach.  There were some minor disagreements, which were all other categories of discrepancies,  and they were observed in about 7% of the tests. If you don't have the rapid phenotyping system, other systems rely on testing for the DNA  and RNA for the presence of organism or resistance genes. With the information regarding the pathogen, hopefully you can make some additional antibiotic decisions.  For example, if you have a gram-positive cocci in clusters and a rapid diagnostic test is  detecting Staph epi, you might be able to make a clinical decision whether that's considered a contaminant.  Similarly, if you have GPCs in pairs and chains and it shows E. faecalis, you can know that it's usually an ampicillin-sensitive organism, which you may be able to de-escalate away  from vancomycin. Vice versa, if there was an E. faecium, then it's likely vancomycin-resistant, and you may need additional antibiotics such as staphylmycin.  Similar inferences can be made if you have an E. coli and you know that your local susceptibility  suggests that dual gram-negative is not necessary, or if you have an albicans, then fluconazole is usually sufficient.  If you have rapid diagnostic platforms that can detect both organism and the presence of resistance genes, then you may be able to make even more inferences.  For example, if you have a Staph aureus with no MEK-A gene, then vancomycin is probably no longer necessary.  Similarly, if you have an E. coli and no CTXM gene, maybe carbapenems are no longer needed.  Of course, these tests can also help us identify pathogens that we're not currently covering and get us started on the correct antibiotics earlier.  For example, if you have one of these carbapenem-resistant genes with an enterobacteriaceae, then maybe something like colistin or Ceftas-AV is warranted.  Likewise, if you have a carbapenem-resistant gene with a pseudomonas, then maybe Ceftola-Xanthazobactam is necessary.  One common thread among all of these rapid diagnostic tests for pathogens and resistance genes is that it doesn't give us the complete picture of the susceptibility results.  We have to make the best use of the information that is available to us and continue to tailor when full susceptibility results are available.  These rapid diagnostic tests have been evaluated in many different studies, and for the most  part, they have all found similar results as long as they were used as part of an antimicrobial stewardship program.  This particular study was by Bookstaver and colleagues, and they implemented a rapid diagnostic  test within antimicrobial stewardship program and evaluated time to de-escalation among patients with gram-negative bloodstream infections before and after implementation.  As you can see from this chart here, looking at combination gram-negative, the use of anti-pseudomonal  beta-lactams, and also carbapenems, time to de-escalation was all decreased with the use of rapid diagnostic tests. In each of these comparisons, the p-value was less than 0.001.  I think it's important to highlight that rapid diagnostic tests is only part of the puzzle when it comes to antimicrobial stewardship, and that by themselves, they're unlikely to  make any significant improvement towards more judicious use of antimicrobials. This study will be covered by Dr. Soros later on, so I won't focus on it, but I do want  to highlight a few things from here. First off, of course, if stopping broad-spectrum antibiotics quicker leads to worse patient  outcomes, then it's likely meaningless, but I think this meta-analysis here can demonstrate that patient outcomes does not suffer.  This forest plot is probably too small to see, but it summarizes the mortality outcomes associated with rapid diagnostic tests.  Majority of the studies were pre-post-implementation studies, very similar to the study by Bookstaver and colleagues in the previous slide.  Overall, mortality was in favor of using rapid diagnostic tests. However, looking closer at the subgroups of those who incorporated rapid diagnostic tests  within an antimicrobial stewardship program versus those that did not, what you observe  is that there was a significant benefit associated with its use when combined with a stewardship program, but not when a stewardship program was not present.  I think this speaks to the fact that having these tests are not enough, but that commissions  us to be committed to using the information, and incorporating them as part of an antimicrobial stewardship program is probably the best way to go about that.  I think this study further illustrates this point. This is a randomized control study on inpatients with positive blood cultures, and it compared  three groups, standard blood culture, multiplex PCR rapid diagnostic tests with templated language, which was used to assist the care team in interpreting the results, or the same  multiplex PCR system implemented with guidance from an antimicrobial stewardship team. This PCR system can detect gram-positive cocci, gram-negatives, Canada, and also selective  genes for gram-positive resistance. It included a little bit over 600 patients. The p-value listed here is comparing all three groups together, and what we can observe is  that there was a significant difference between groups when it comes to both time to escalation and time to de-escalation. However, we can make a couple of additional observations.  In all of the groups, time to escalation from new information was quicker than time to de-escalation,  which I think once again speaks to clinicians' natural inclination to want to provide more antibiotics as opposed to stop antibiotics.  The authors also performed post hoc pairwise comparisons, and in both time to de-escalation  and time to escalation, only the group assigned to stewardship program was significantly different to control.  Looking closer, though, for escalation, there was really very little difference between the stewardship or the templated language group.  However, when looking at de-escalation, the stewardship group was not only significantly better than control, it was also better than the group that had template language.  I think in conclusion, this study demonstrates that implementing rapid diagnostic tests is  not enough, and that particularly for de-escalation, combining it with some sort of formal stewardship effort is necessary.  So in conclusion, I would say that antibiotic use and resistance continues to increase. Surveillance cultures using rapid diagnostics may be helpful in empiric antibiotic selection  and or de-escalation, and that rapid diagnostics, when clinical cultures are positive, can give us earlier information.  However, having that information is not enough, and that early information is probably best paired with some sort of stewardship efforts to get the most benefit out of it.  Thank you for your time and your attention. Hi and welcome to this talk on evaluating rapid diagnostic testing for critically ill patients.  I'm Lauren Source, Founder's Board Nurse Scientist at Ann and Robert H. Lurie Children's Hospital  of Chicago, Northwestern University, Feinberg School of Medicine, and I'd like to thank the organizing committee for inviting me to give this presentation.  Of note, I have no disclosures regarding the content of this talk. So we're going to talk about and evaluate the evidence surrounding the use of rapid  diagnostic testing in critically ill patients in this session. So first, let's start out by defining rapid diagnostic testing. What is it?  So it's testing for a condition or presence of a disease with a rapid time to result. But the question arises, what does rapid really mean? Is it five minutes to result time?  Is it 60 minutes? Is it two hours? Or is it 12 hours? Or is it something completely different? Is it a percent decrease in time compared to the standard test result?  Or is it a percent decrease in time compared to the standard test that leads to a change in patient outcome?  So while we're still thinking about that, it's important to know that rapid diagnostic testing can be performed by a variety of professionals and lay people.  There are some pros and cons to using this type of testing. Pro is easy to use usually with rapid results, again, defined by a different definition for each test.  And there may be some need for limited infrastructure instrumentation. But on the con side, the interpretation of the results may be subjective.  There may be limited sensitivity compared to the gold standard and likely a need for quality of control.  Another pro may be a substantial need for infrastructure and implementation, depending on the test.  So it then begs the question if rapid diagnostic testing and point of care testing are the same. Certainly, rapid diagnostic testing can be point of care testing, for example, blood  gas analysis. And point of care testing can be a rapidly diagnostic test, for example, blood gas analysis.  Once you obtain your blood gas, you know whether or not your patient has an acidemia and frankly what type.  Some common applications of point of care testing are glucose, commonly used in lay population as well as pregnancy tests in that same population.  And then used variably through the lay and the medical population, urine dips, fat grab, blood for stool, blood gases, bedside electrolytes, lactate, and the like.  So I would pose this question for your consideration. What about lab tests that get sent down to the lab?  For example, chemistries, hepatic panels, C-reactive protein, procalcitonin. How did you change your practice and did you see a change in outcomes when some of these  tests became rapid? Because all these tests weren't rapid. So let's look at the first study that's examining the mask two phase process, MALDI, time of  flight, short incubation identification from blood cultures and comparing 396 adults published by Delfort and colleagues in 2017.  They did a retrospective analysis where they compared standard blood culture in 2012 with the MALDI time of flight process in 2013 as the second group to the third group with rapid  short incubation protocol. That rapid short incubation protocol used three hours of blood culture on agar then leading to doing MALDI time of flight.  The outcomes they looked for were time to appropriate antibiotics, length of stay, and mortality risk.  So looking at this slide for all comers, looking at days in the hospital on the left and time  into antibiotic identification and identification and administration in days on the right, when you look at the conventional group or group one, you can see a consistent decrease in  days in the hospital between group one and group two and a further decrease in group three, which was the short incubation protocol.  Overall, the length of stay was decreased by all comers in this group by 4.54 days. And that was a statistically significant difference.  When we look at a subgroup of patients who needed optimizing antibiotics after the identification of the blood culture pathogen, again, you can see a decrease in length of stay from  the conventional group to the middle group or the MALDI time of flight to the short incubation protocol group.  And again, that change is statistically significant in a reduction of nearly seven days. When we look at all comers in each group, and you can see indicated by this hash mark  box in each of the groups, that this is the group that was already on the correct empirical therapy, and this is the group that needed a change.  You can see these lines begin to converge and resulting in an overall improvement in the care of the patients received when they had the rapid incubation protocol.  Duffport also did the same exact methodology, but he did it in children. This subgroup was 92 children, again, retrospective analysis and comparing standard blood culture  to the MALDI time of flight and the rapid short incubation protocols, again, the same strategy as in the previously discussed study.  The outcomes here were timing of appropriate antibiotics in patient outcomes. It's important to note that the groups all had similar demographics.  What's most important to point out is this substantially significant result in time to identification, regardless of which groups were being compared.  That was consistent across all comparisons for all groups. What was not statistically significant across all groups was length of hospital stay.  And that may be a result of the low numbers of patients within each group. However, time to final identification was positive when comparing the rapid incubation  protocol with conventional, as well as was time to antibiotic change in the MALDI time of flight identification and the rapid identification protocol.  When we're looking at the impact of length of stay in the different subgroups, you can see that the greatest change came, continued to come down from the conventional to the  MALDI-TOF group to the short incubation period, with a 75% improvement in length of stay from  44.4 days to 10.94 days between the conventional blood culture group and the rapid short incubation group.  When we looked at the group needing antibiotic change to the group that was on the proper  empirical antibiotics to start with, the biggest change we see is in the conventional identification group with the biggest difference. Otherwise, the difference was not seen.  So let's move on to a pre-post-intervention study published by Verican and colleagues in 2019, looking at rapid molecular identification with positive blood cultures.  In this group, they specifically studied 287 adult ICU patients. In the pre-intervention group, they looked at blood culture and gram stain with the MALDI time of flight test.  The intervention group used blood culture and then molecular testing using the BioFire molecular test. The outcome was time to appropriate antimicrobial treatment.  For patients who had the molecular test done, once that molecular test was done within one hour of testing positive blood culture, those positive cultures were immediately called  up to the ICU team. However, if a positive blood culture was found and the molecular testing identified no bacteria  or no organism, then no call was made. So let's look at the results. To start with, you should know that in the pre-intervention group, patients were older,  had more cardiovascular disease, and had more chronic lung disease compared to the post-intervention group.  However, when we look at P0, so that's the pre-intervention, this is the time of a positive blood culture.  Time starts ticking, and it takes 14.41 hours to get the results of the MALDI time of flight result, and then a near-immediate optimization of antibiotic therapy results.  In time two, or the post-intervention, the blood culture is positive. Before one hour, we have the results of the molecular test, followed by a call to the  ICU, and a resultant optimization of antibiotics. You can see the statistically significant improvement in the time of identification of the positive blood.  So let's look at a modeling study done in 283 adults with respiratory symptoms, looking at molecular point-of-care testing on clinical management.  So what this group of researchers did was they used prospective data collection for a cost analysis study where they collected respiratory samples on patients and tested  them, and then saved a portion of the sample for the later part in the modeling part of this study.  So as you can see from the screen, that is the data that they collected, and they proposed two scenarios.  Scenario one was point-of-care testing and the impact on time to diagnosis, and point-of-care testing impact on discharge as scenario number two.  This was the model that they used in their study, and they assumed within this model that the point-of-care testing would provide test results and diagnosis within one hour  of admission. As you can see here, patients who start either in the outpatient arena or in the emergency  department can go to a variety of different places as they travel to discharge or death. So let's look at what happened. In this screen, we can see this is where patients went.  You can see on the solid line of red and the solid line of blue, as well as the hash line of red, that patients with unknown respiratory status were either admitted to the ward in  isolation or not in isolation, or patients in the hash red line, their respiratory infectious status was known, but they were inappropriately isolated.  When we look at scenario one, identified here as B, for time to diagnosis, we can see that we no longer see the red and the blue solid lines and the red hash lines, because patients  were put where they needed to be with known diagnosis within one hour. Of note, these patients had a discharge of about 70% of them by 200 hours, which is consistent  with what happened initially in real time. When we look at time to discharge, we can see that when we modeled for it, that again,  patients were identified within one hour and went to the units they belong to, either in or not in isolation, and nearly all patients were discharged by the 200th hour compared  to 70% in real life. When we look at the cost in money, this was done in Europe, so the cost was in euros,  we can see that there's no information identified for ICU time, so perhaps ICU cost is the same as ward cost.  But you can see, compared to usual care, scenario one and scenario two, that ultimately it costs less if you have point of care molecular testing, the care provided.  Let's look at a quasi-experimental study by Wave and colleagues published in 2019, looking at over 4,000 adults, which included 2,000 control patients.  Their study was non-randomized, quasi-experimental, comparing the standard multiplex PCR with rapid PCR and a control group.  Pre-implementation, patients received the standard multiplex PCR, and post-implementation, they received the rapid PCR.  Outcomes were length of stay, and then secondarily, microbiology tests. It's important to note that there were differences between the groups with the intervention groups  coming more from the emergency department, more went to the ICU, and more were diagnosed with influenza and pneumonia.  When we looked at the intervention group overall, we know that the median length of stay decreased  significantly, and for those patients who remained in the hospital when their test was resulted, we can see regardless of positive or negative tests, the length of stay was  significantly reduced. When we look at how much secondary testing was done, we can see in group A, which were  all the all comers positive, we could see, or regardless of test results, we can see that there's significantly less ordering of serologies and sputum and blood culture.  In group B, those with just positive test results, we can see there was only significantly reduced ordering for serologies and sputum culture.  In group C, those with negative tests, it was still less ordering of bacterial and viral serology as well as sputum cultures and blood cultures, and group D, when we compared the  control arms of the pre and the post, we can see that there was less ordering, only statistically significant in blood cultures.  Now let's look at studies that were done looking at a new method, a polymerase chain reaction followed by electrospray ionization mass spectrometry done on 616 specimens from 529  critically ill adults in six European countries. This was a multi-center observational trial where they compared standard microbiology testing with this culture-independent testing.  In the blood cultures, we can see that there were 228 positive samples using the novel method compared to 68 using the standard blood culture method.  We can see that there was a negative predictive value of 97% and a sensitivity of 81%. When we look at the lower respiratory tract and the sterile fluid tissue, we can see that  the sensitivity and specificity are similar, but the negative predictive value of both tests are different from the bloodstream testing.  As part of this study, these results were given to a panel of experts, and the panel of experts said that had they had this testing, over slightly one-third of them, they would  either alter, institute, or cease antibiotic therapy earlier based on these results. Let's look at a systematic review that was published by Timbrook and colleagues in 2017  where they looked at 31 studies. Most represented were adult-based studies, six were international, and it included sites with and without antimicrobial stewardship programs.  However, it's important to note that not all of the antimicrobial programs that were reflected in these studies were actually in real time.  They compared conventional blood culture with PCR or MALDI time of flight, PNA-FISH with blood culture gram-negative nanotechnology, blood culture gram-positive nanotechnology,  or a combination of both of those. The outcomes of interest were 30-day all-cause mortality risk, time to effective therapy, and length of stay.  Looking at the subgroup of studies that had an antimicrobial stewardship program, we can see that running molecular rapid diagnostic tests are favored in the reduction of all-cause  mortality. These studies had very little heterogeneity. When we look at the subgroup of papers that did not have antimicrobial stewardship programs,  there were fewer of them, and we can see that there was not a statistically significant difference to favor molecular rapid diagnostic tests to reduce all-cause 30-day mortality.  There was some heterogeneity within these papers, but when we look overall, taking both studies and subgroups, we can see that rapid diagnostic testing overall reduces all-cause  mortality in these patients. When we look at time to effective therapy, we can see there are a variety of subgroups analyzed in this porous plot.  However, when we come to the bottom, you can see that the time to effective therapy is no different, taking all comers, acknowledging this 34% I-squared number, so a little bit  of heterogeneity. For length of stay, there were not that many studies, and they had large and varied heterogeneity.  However, the p-value in this porous plot is significant that molecular rapid diagnostic testing does favor reduction in length of stay.  So a brief word on SARS-CoV-2 and rapid diagnostic testing. The type of testing is done and is rapid depending on what is actually being tested.  The sample source will all dictate the time to result. For those of you looking for more evidence and more support and information on SARS-CoV-2,  PubMed has done a very beautiful job separating all of that out for us, and when I looked  for diagnostic testing in this screen, I found that there were over 13,000 papers published, and you can see how frequently. So let's end with this.  The Surviving Sepsis Campaign was organized of 16 international experts convened to identify  research priorities in sepsis and septic shock because known time sensitivity to diagnosing sepsis leads to decreased mortality, length of stay, hospital costs, and antibiotic use.  However, there are gaps. While we await days for results and tests, difficulties with sampling, and IDing resistant pathogens once they're identified.  So led by Dr. Coopersmith, the group came up with these six research priorities, and in these six research priorities that were not in order of importance is, should rapid  diagnostic tests be implemented in clinical practice? So given the dearth of literature on the critical ill patient on rapid diagnostic testing, this  remains a research area of focus for our community. Thank you for your kind attention.

Video Summary

The utility of rapid diagnostic testing in critically ill patients is a topic of interest and ongoing research. Rapid diagnostic testing refers to tests that provide results quickly, usually within a few hours or less. These tests can help healthcare professionals make quicker and more accurate diagnoses, leading to faster initiation of appropriate treatment. In this video, several studies were discussed to evaluate the impact and effectiveness of rapid diagnostic testing in critical care settings. One study focused on the use of MALDI-TOF mass spectrometry to identify pathogens in positive blood cultures. The study found that rapid identification of pathogens led to shorter hospital stays and improved patient outcomes. Another study evaluated the impact of rapid molecular testing on the management of respiratory infections. The study showed that rapid testing resulted in faster diagnosis and improved patient care. Additionally, a systematic review of 31 studies found that rapid diagnostic testing reduced all-cause mortality and length of hospital stay in critically ill patients. Overall, the evidence suggests that rapid diagnostic testing can improve patient outcomes and streamline care in critical care settings. It is an area of ongoing research and continues to be a research priority in the field of critical care medicine. By implementing rapid diagnostic testing, healthcare professionals can make quicker and more accurate diagnoses, leading to improved patient outcomes and more efficient use of healthcare resources.

Asset Subtitle

Procedures, Infection, 2021

Asset Caption

The Society of Critical Care Medicine's Critical Care Congress features internationally renowned faculty and content sessions highlighting the most up-to-date, evidence-based developments in critical care medicine. This is a presentation from the 2021 Critical Care Congress held virtually from January 31-February 12, 2021.

Learning Objectives

Distinguish the differences among rapid diagnostic tests (RDTs) for bacterial, viral, and fungal infections
Evaluate the evidence surrounding the use of RDTs in critically ill patients
Assess the use of RDTs in the management of infections in critically ill patients for antimicrobial stewardship

The Utility of Rapid Biomarkers Sameer Kadri Antimicrobial Stewardship Utilizing Rapid Diagnostics Simon W. Lam Evaluating Rapid Diagnostic Testing for Critically Ill Patients Lauren R. Sorce

Meta Tag

Content Type Presentation

Knowledge Area Procedures

Knowledge Area Infection

Knowledge Level Foundational

Knowledge Level Intermediate

Knowledge Level Advanced

Membership Level Select

Tag Point of Care

Tag Antibiotics

Year 2021

Keywords

rapid diagnostic testing

critically ill patients

ongoing research

quicker diagnoses

improved patient outcomes

pathogen identification

shorter hospital stays

faster diagnosis

patient care

critical care settings

	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.