for inviting me. It's become a little bit of a pattern, I think, for organizers to schedule me as the next-to-last speaker, and this recently happened at the CHESS meeting, and when the last speaker ended up with only five minutes of time for a 15-minute talk, I had a few four-letter words thrown in my direction, so I'll try to avoid the same outcome in this case, and therefore, we'll try to stick as much and as faithfully to simply guideline data as possible. So, as you heard, of course, this is the part about invasive candidiasis. The guidelines that we will survey will have been published in the years 2019 and 2021, which I think is apropos of the fact that this is a kind of a reunion conference of sorts, and you'll appreciate that guidelines published in these years could have easily been sort of subsumed in the chaos and tumult of the COVID pandemic and therefore may have not gotten necessarily their due attention. This is just a brief bio slide. Up on top is Westchester Medical Center. Below is New York Medical College. They're neighbors on the same campus in Valhalla, New York, just north of New York City. I have no disclosures. I will be mentioning or showing some unpublished data from our American Thoracic Society Working Group, but there is no conflict of interest inherent in any of that material. So, a few objectives. The emphasis, just to be clear, will be on non-neutropenic ICU patients with invasive candidiasis. I think the neutropenics are a bit of a different animal, and so they won't really be the focus of this discussion. We'll start with a brief background on invasive candidiasis in the ICU, so I'll abbreviate invasive candidiasis as IC going forward. We'll survey the guideline recommendations on the choice of the antifungal in these cases in the ICU, guideline recommendations on diagnostic biomarker use, and guideline recommendations on, quote-unquote, untargeted therapy, which is, as I'll mention later, therapy administered in the absence of culture confirmation of infection. So, what is the scope of the problem? I'd like to make the argument that it's a very substantial scope. Invasive candidiasis is the most common invasive mycosis in the ICU. I would argue that it's unique among invasive mycoses in that it is most commonly diagnosed in blood culture, and that's what, of course, we term candidemia. This is very different from other mycoses you might imagine, such as aspergillus infection or the endemics. Its crude mortality has been shown to exceed 60%, especially in certain situations, like patients in the medical ICU, as shown in a study from our institution, and if septic shock is present, as was shown in another U.S. study. Interestingly, these U.S. data tend to be somewhat higher than the European numbers, and the reasons for this discrepancy aren't really well elucidated. And as I'll try to impress upon you in the next slide, it's a very prominent pathogen in ICU-acquired bloodstream infection. So I promised very little deviation from the survey of the guidelines themselves, but I'd like to spend just a couple of minutes on this particular slide, dissecting it, even though it's really not part of the main thrust of the conversation. But it comes from a very timely study published in Critical Care Medicine just very recently, which is, and it's a big U.S. sample, 151,000 ICU patients at 85 hospitals. And I'll turn your attention first to the graph on the left, where you have on the x-axis number of pathogens implicated in ICU-BSI, and on the y-axis the absolute numbers that different pathogens contribute. And you see the blue arrows point to the non-candida, non-Albicans candida, rather, and the Albicans candida, respectively, which in these absolute terms, in this overall graph, do not appear to make a very impressive contribution. Just on a parenthetical note, it's interesting that in this large U.S. sample, non-Albicans candida numerically exceeded Albicans candida, which we've seen in our own institution, but has a lot of geographic variability. So interesting that in this U.S. sample, non-Albicans candida were numerically more represented, and that, of course, includes your resistant species, such as Labrada and the very currently feared Candida auris. But really, I think the bigger lesson comes from the graph on the right, in which the authors, again, have the same pathogens on the x-axis. On the y-axis is now percentage of patients with BSI in the ICU. But what's particularly noteworthy is that now each pathogen is broken down into blue and orange bars. So the blue bars depict bloodstream infections that were present on admission to the ICU, or identified almost immediately thereafter. So essentially, bloodstream infections with which the patients were admitted to the ICU. And the orange bars represent ICU-acquired bloodstream infections, so those that were identified on day three or later of ICU stay. And when you look at those orange bars, you start to notice that the non-Albicans and Albicans Candida all of a sudden are making a much more prominent contribution to these percentages. And I think it's even more telling when I draw a separate bar in which I combine the non-Albicans and Albicans species, which then now amounts to a 16% contribution and makes Candida actually the commonest cause of ICU-acquired bloodstream infection among all these pathogens. And I think this may be surprising to quite a few folks, and I've had my ID colleagues really express a lot of surprise by this, but in reality, this is now part of an emerging pattern of data regarding the role of Candida and its prominent role, as I'm showing you here. So we'll dive right into the guidelines and pose sort of the easiest question we have to contend with, which is the choice of antifungal agent in the intensive care unit. And for this, we will first refer to a document from Intensive Care Medicine published in 2019. This is an effort by our European colleagues, both from the European Society of Intensive Care Medicine and European Society of Clinical Microbiology and Infectious Disease. And you'll see this guideline reappear a few times during my talk, so I'll just refer to it generically as our European colleagues for simplicity. And not perhaps surprising, surprisingly, the consensus statement on choice of agent stated that the panel recommends that echinocandins should be used as the first treatment option in critically ill patients with septic shock and multi-organ failure with invasive candidiasis. And as you see there, it was graded as a weak recommendation based on low-quality evidence. I'd like for you to keep that thought in mind as we view the subsequent slide. There's a second bullet point here having to do with less severely ill ICU patients in which fluconazole was viewed as an acceptable alternative but I don't think those are really the patients that we're most fearful of in high-acuity ICUs. And so in 2021, our Australasian colleagues actually published a guideline document that in part addressed this very same question. And of course, again, the echinocandins came out as the agent of choice for the treatment of IC in the ICU. But what you will notice is that in this case, now two years later, they assigned this recommendation the highest strength of recommendation and the highest quality of evidence grade. And so what happened in the intervening two years? You'll recall that this, sorry, that this was a weak recommendation, low-quality evidence and turned into a very strong recommendation based on high-quality evidence. And that is the interval publication of important studies establishing now more firmly the role of echinocandins as the drug of choice for this in the intensive care unit. And as I, you know, from experience, I can tell you that it takes many years to develop a guideline document. And so it seems like only two years separate these two guidelines, but in reality, the literature reviews may have been separated by a longer time than that. So pivoting to biomarkers, in this day and age, although there are a number of biomarkers out there, really the conversation still centers on beta-D-glucan because it is the most widely available and the most thoroughly studied. It is the cell, it is a cell wall component of most fungi, including candida, and some, including some of my colleagues, refer to it as sort of the ESR of fungal infection. In other words, the reliance on it is to serve as a red flag to alert us to the possibility of the presence of a fungal infection, not necessarily a specific fungal infection. And so this question became one of the questions addressed in the 2019 guideline document issued by a group of us on behalf of the American Thoracic Society. And so on the biomarker side, we looked at 10 studies that were pooled, 10 relevant studies. There was very robust outcome ascertainment or disease ascertainment in that eight out of the 10 studies included patients who only had proven invasive candidiasis. So that was a very robust gold standard in the majority of the studies. And as a red flag, of course, we want to see, we want to focus on the sensitivity of something that is supposed to be a red flag, that's supposed to alert us to the possible presence of a fungal infection. And when we looked at these 10 studies in aggregate, this was the sensitivity, the pooled sensitivity number that we derived. And so what was the question that we asked framing this data analysis? The question was phrased as follows. In critically ill patients with suspected invasive candidiasis, is the beta-D-glucan assay alone, and I emphasize alone, sufficient for diagnostic decision-making? And the answer, in critically ill patients in whom there is clinical concern for invasive candidiasis, we suggest against relying solely on results of serum beta-D-glucan for diagnostic decision-making. And this was classified as a conditional recommendation based on low-quality evidence. So what was behind the rationale for this recommendation? We were, of course, discouraged by the sensitivity of 81% for something that's supposed to be a red flag, which means that there was a false negative rate of 19% for what is a very serious infection. And so what I think is helpful to visualize this is if one takes the positive likelihood ratios and the ratio and the negative likelihood ratio that was derived from this pooled analysis and applies it to a patient with 50-50 pretest probability of invasive candidiasis. So a patient in whom you're on the fence about the diagnosis. And so if you were to apply the positive ratio to somebody with positive beta-D-glucan level, the post-test probability goes up to 68%. And if you were to take the same patient and apply the negative likelihood ratio with a negative result, the post-test probability drops to 24%. So the test does move the needle, but arguably doesn't move it enough. And an important point to mention here is that this recommendation was issued on the basis of largely indirect evidence because nine out of the 10 studies actually relied on acquisition of beta-D-glucan tests based on a schedule, based on timing, as opposed to clinical suspicion. And really it was clinical suspicion that was part of our question. And so what the silver lining of all of this is is that the one study that did not use timing and did not use a schedule for the beta-D-glucan essays, but rather used clinical suspicion, actually resulted in substantially better performance characteristics of beta-D-glucan than our overall pooled analysis showed. So the third and last sort of component of the guideline survey will involve so-called untargeted therapy. And so when we refer to untargeted therapy, what we mean is that, as I briefly mentioned earlier, it is that therapy which is initiated in the absence of diagnostic culture results. And untargeted therapy comes in three flavors. The first flavor is prophylaxis. That is antifungal therapy is administered for quote-unquote at-risk ICU patients. And the definition of who qualifies for being at risk is, of course, a matter of debate. The second flavor is preemptive therapy. That is therapy initiated in the event of positive biomarker surveillance. And so whatever the schema for biomarker surveillance might be, that's also variable. Of course, these first two groups are patients who don't have clinically evident infection. And the third flavor, which is perhaps the most near and dear to the hearts of practicing intensivists, is empirical therapy, which means antifungal therapy administered for suspected candida sepsis, but of course without microbiological confirmation. And so first turning to prophylaxis, I show you an example of one of what are several studies that have been performed looking at prophylaxis of ICU patients. This happens to be prophylaxis in the surgical ICU, fairly classical trial from a number of years ago. The reason I show it is that its results are fairly standard for these trials in general in that they show no difference in survival, but they do show a decrease in incidence of invasive candidiasis. And so our European colleagues asked this question as question number three in their 2019 intensive care medicine document. Should antifungal prophylaxis be used in critically ill patients? The answer, as a consensus statement, the panel recommends against the routine and universal administration of antifungal prophylaxis in critically ill patients, which was deemed to be a weak recommendation on moderate quality evidence. And just to put somewhat of a more graphic, I guess some graphics around this, this is part of the unpublished data that I alluded to, a meta-analysis that we have performed as part of an evolving American Thoracic Society guideline project and now having to do with fungal therapeutics. And so as you can see here, this is a pooled analysis of prophylaxis with mortality as the endpoint. And in the forest plot, you can see that the pooled estimate centers right on the line of no difference. Preemptive therapy, there are no studies directly addressing this flavor of untargeted therapy. I do include here two studies on the subject, showing, for example, that it is feasible to implement, that it lowers beta-D-glucan levels, that it is perhaps economically feasible, clinically feasible, et cetera, but no study directly addresses this question. And so our European colleagues made this question number four in their document, should preemptive therapy be used in critically ill patients? And the answer, as part of a consensus statement, was that the panel does not recommend the use of preemptive antifungal therapy in critically ill patients. Again, not surprisingly weak recommendation, low quality evidence, almost absence of evidence in reality. And so again, near and dear to our hearts, the often-encountered question in the ICU of empirical therapy. This is, of course, the patient with clinical infection, shock, potentially on antimicrobials, or on antibacterials, rather. And the question is, should empirical antifungal therapy be started? And when we talk about antifungal therapy in that context, we're really only talking about invasive candidiasis, of course. So our European colleagues made this question number five, which patients should receive, slightly different wording, which patients should receive empirical antifungal treatment? The answer is a consensus statement. The panel suggests that empirical antifungal therapy might be considered only in patients with septic shock and multi-organ failure who have more than one extra digestive site with proven candida species colonization. And in this context, I remind you of something called the candida score, somewhat parenthetical to this, but nonetheless, I think, instructive to look at. So candida score is a clinical prediction rule that was developed some number of years ago now to allow, essentially, to allow a clinical determination of the likelihood that a given ICU patient is suffering from invasive candidiasis. And so if the candida score were to be applied to the patient described in this consensus statement, they would receive two points for being in septic shock, and they would receive one point for multifocal candida colonization. So their candida score would, at a minimum, be three. And the reason why this is relevant is shown in the little excerpt at the bottom, and that is that a score of three in the candida score is a breakpoint in that clinical prediction rule, such that patients with scores of three and higher are at multifold higher risk of having invasive candidiasis than those who have scores less than three. And so, again, unpublished data to this point, but part of our guideline, ongoing guideline effort was a meta-analysis of trials of empirical therapy. And much the same, you see that the summary pooled estimate is graphically centered right on the line of no difference. So to conclude, try to keep us hopefully on good time, candida species are a very prominent cause of ICU-acquired bloodstream infection. Hopefully, my taking a few minutes on those bar graphs has convinced you of this. Candidacandids are the drug class of choice for invasive candidiasis in the critically ill. Beta-D-glucan has a substantial false negative rate in ICU patients with invasive candidiasis, and not to be overlooked, an even higher false positive rate. Judicious prophylaxis may reduce invasive candidiasis, but no real data regarding a positive impact on mortality. Likewise, a mortality benefit has not been shown for empirical therapy. Thank you, and then we'll be available for questions. Thank you.

invasive candidiasis

ICU patients

echinocandins

beta-D-glucan

empirical antifungal therapy