Hello everyone, I'd like to thank the Congress Committee for inviting me to speak to you on the role of biomarkers in the detection of infection in sepsis. As a case in point, I will be focusing on procalcitonin as it is the bacterial infection biomarker that is the most extensively studied, most often used, and also remains rather controversial. I have no financial disclosures. I serve in the IDSA Sepsis Task Force and CMS Sepsis Technical Expert Panel, and I also serve as an Associate Editor for Critical Care Medicine. I'll be focusing on the epidemiology of sepsis, antibiotic use, and procalcitonin, touching on the significance of procalcitonin in COVID-19, and then taking sort of a deep dive into the evidence behind starting and stopping antibiotics in respiratory tract infections, as well as critical illness and sepsis. And in the end, provide recommendations for clinicians, as well as for future trials. In 2007, Jean-Louis Vincent published an important global point prevalence survey of 1,265 ICUs, where he asked the question, on any given day in an ICU, what proportion of patients are on antibiotics? And he found it to be 71%. Fast forward 10 years, this point prevalence study was repeated, and a lot changed in that decade, especially around situational awareness of antibiotic resistance and the importance of stewardship. And even after 10 years, the proportion still remains at 70%. This diagram represents the three concentric milieu of antibiotic resistance, the inner sphere representing the bedside, the middle sphere representing the hospital, and the outer sphere representing the environment. And we often focus on managing the patient, often ignoring the distant threat sort of for antibiotic resistance. But unfortunately, it is multiple decisions that are taken within that inner sphere over many, many years that has catapulted the antibiotic resistance crisis into what we know of it today. On the other hand, the price of missing sepsis is high. As you can see here, the annual burden of deaths related to sepsis in the US is 270,000, and globally is 5.3 million. Despite years of investigations into the pathobiology of sepsis, we still do not have a diagnostic gold standard for sepsis. Even among intensivists who see sepsis every day, there is significantly poor interrater agreement in terms of who has and who doesn't have sepsis. There are also layers of uncertainty in the management of sepsis, such as, should I treat empirically? What's the bug? How long should I treat? How long should I use broad-spectrum agents? In this US study of culture-positive community-onset sepsis, the burden of antibiotic use, broad-spectrum antibiotic use was very high, but at the same time, the proportion of resistant pathogens was quite moderate. Therapeutic progress is stymied in part by lack of rapid and specific diagnoses. It makes no sense to use 21st century technology to develop drugs targeted at specific infections whose diagnoses are delayed by 19th century methods. A lot of sepsis still remains culture-negative, about 30-58%, and studies have shown that culture-negative sepsis is just as bad as culture-positive sepsis. In culture-negative serious infection, early discontinuation of antibiotics is infrequent and variable, as shown in this multicenter study, because essentially, management of culture-negative sepsis is a data-free zone. All of this has made very intriguing the concept of biomarkers in the management of sepsis. Biomarkers broadly defined are markers of a biological process or state, but before integrating biomarkers into diagnosis and management, you have to ask yourself the question, is it going to inform me anything beyond conventional clinical judgment? Will it lead to improved patient outcomes, or will it save me money? Several biomarkers have been tested over the years, but the only one that's made it into routine guidelines is procalcitonin. Procalcitonin has been FDA-approved in the US for risk assessment of sepsis on ICU admission, to prognosticate severe sepsis using serial measurements, as well as in the starting and stopping of antibiotics in lower respiratory tract infections, and in the stoppage of antibiotics in sepsis. Proclacitonin has some unique characteristics that make it desirable as a biomarker. So in comparison to CRP, TNF-alpha, IL-6, PCT has a pretty rapid onset of two to four hours from the onset of infection, and then it peaks at about 12 to 24 hours. PCT outperforms other biomarkers in discriminating bacterial infection. In a meta-analysis of 30 studies, 3,244 critically ill patients, 57% of whom were septic, the pool sensitivity was 0.77, pool specificity of procalcitonin was 0.79, and procalcitonin had an area under the curve of 0.85. So it's not a perfect marker, but a perfect marker does not exist. There are potential false positives, however. Proclacitonin can be elevated in non-infectious conditions, such as trauma, surgery, hyperthermia, CKD, neoplasms, as well as certain non-bacterial infections, such as malaria, dengue, and now severe COVID-19. PCT varies between gram-positive and gram-negative infections. It's higher for gram-negative infections, but I think this information is of unclear clinical significance. There is significant biological variability in procalcitonin levels, as shown in this large study of 42,000 patients with bloodstream infection, a fifth of whom met sepsis criteria. Interestingly, we found that in the US, in this study, 10% of patients with suspected bloodstream infection slash sepsis were tested for procalcitonin, and the median level was 4.89. But importantly, severity of illness was a much larger driver of high PCT levels than, for instance, what pathogen or what site the infection was at. While it is well known that with every hour gone without antibiotics in septic shock, mortality goes up by about 7%, this recent study shows that for even patients with sepsis without shock, their risk of progression to septic shock increases by about 4% for each hour of delayed antibiotics. The surviving sepsis campaign 2021 guideline update also tells us that antibiotics for suspected sepsis without shock should be based on likelihood of infection, and this is an update from the prior guideline. If the likelihood of sepsis is high, give antibiotics within one hour. If the likelihood of infection is low, perhaps defer antimicrobials while closely monitoring and rapidly investigate to find alternative reasons for the presentation, and if infection concern still persists, try to give the antibiotics within three hours. And one wonders, with this new guideline recommendation based, basing antibiotic use on likelihood of infection, is there a role of procalcitonin in helping with this likelihood? Can procalcitonin help with prognostication? This study, the MOSES study, was a multicenter prospective cohort study of 13 US EDs of sepsis patients in whom procalcitonin was measured daily, and they found that if procalcitonin did not decrease by more than 80% over the five-day period, that predicted a higher mortality. But I wonder what to do with that information, and perhaps procalcitonin alone cannot be used to prognosticate patients. Also, can we escalate therapy based on high procalcitonin levels? This study, the PASS study, was a randomized clinical trial from nine ICUs in Denmark, and they had the intervention of increasing diagnostic studies and mandated antibiotic escalation if procalcitonin levels remained high. And this was compared to standard of care. And the key finding was that while there was no difference in survival, there was greater acute kidney injury, dialysis needs, and mechanical ventilation needs in the procalcitonin group. And as such, procalcitonin-guided escalation in the ICU cannot be recommended at this time. In this study of 60 hospitals, 42,000 patients with bacteremia, the false negative rate for admission procalcitonin was 16%, and varied considerably by pathogen, and was more common in bacteremia without sepsis. The 2021 guideline update for the surviving sepsis campaign recommends for adults with a suspected sepsis and septic shock, they recommend against using procalcitonin for deciding whether to initiate antibiotics. Now let's see what the evidence is around procalcitonin-guided starting and stopping of antibiotics in respiratory tract infections. In a 2017 Cochrane review, Schutz and colleagues showed that procalcitonin-guided antibiotic use appears to be safe in acute respiratory tract infections. They showed lower antibiotic exposure, lower treatment failures, and even a survival benefit. Importantly, the impact of initiation versus discontinuation across these studies is inextricable. And the survival benefit was largely driven by the SAPS trial by Dejong and colleagues, which was a pure discontinuation study. So as such, this sort of collective finding is difficult to generalize solely to the process of antibiotic initiation. And as such, insufficient evidence exists to recommend procalcitonin for antibiotic initiation in community-acquired pneumonia. The IDSA and ATS guidelines for community-acquired pneumonia do not recommend procalcitonin levels to determine antibiotic initiation. The ATS-IDSA guidelines for hospital-acquired and ventilator-associated pneumonia also do not recommend procalcitonin for initiation of therapy. However, they do suggest procalcitonin and clinical criteria could be potentially used for antibiotic discontinuation. Now let's look at, in more detail, procalcitonin antibiotic discontinuation in sepsis. The world of infectious disease is moving towards shorter is better on the basis of several studies that have emerged that have shown sort of equivalent outcome in short versus long courses. However, this interesting systematic review and meta-analysis of hospitalized patients comparing shorter versus longer courses, while they showed comparable efficacy across a variety of outcomes, the important thing to take home here for us is that all these studies predominantly were among ward patients. So as such, is it easy to generalize to our septic patients? Not quite. In fact, higher initial severity and radiologic burden is linked to pneumonia treatment failures in this study. I'd like to see more studies like this conducted in the era of shorter courses of therapy to see, in fact, in the real world, if shorter courses are, in fact, detrimental in serious infections and in sepsis. The surviving sepsis guidelines do recommend shorter over longer courses. However, this is a weak recommendation because there really is insufficient data to extrapolate shorter courses to septic patients. How about procalcitonin-guided discontinuation? Well, 13 meta-analyses on this topic have been done that have essentially shown, you know, decreased antibiotic exposure and length of antibiotic use without worsening outcomes. In fact, more recently, they have even shown a survival benefit. And for acute care providers, survival benefit is a bit of a game changer in terms of interest in use of procalcitonin. But to sort of dig deeper and understand the basis of this, we did yet another meta-analysis. This work was led by Dominic Pepper and is published in CHEST. We looked at 16 RCTs, 5,000 patients. These were all ICU patients from multiple countries, and looked at the comparison of procalcitonin-guided discontinuation versus standard of care, looking at a primary outcome of mortality and a secondary outcome of antimicrobial duration and length of stay. Some notable characteristics of these studies. Adherence was variable. Five RCTs had high and four had low adherence. Seven RCTs were industry-funded, and eight RCTs used both procalcitonin and C-reactive protein in the intervention group. However, one has to focus on the quality of the evidence, and all these studies had a high risk of performance bias. And that's because you cannot blind providers to a procalcitonin algorithm. So this high risk of bias was essentially unavoidable. The study did show survival benefit associated with procalcitonin algorithm, but there are some caveats. There was no survival benefit associated with studies limited to those in patients with sepsis, studies in which procalcitonin algorithm adherence was very high, non-industry-sponsored studies, and in studies in which procalcitonin was used exclusively. Most of the survival benefit in the meta-analysis can be attributed to this study, the SAP study by Dejong and colleagues. 1,575 critically ill patients were randomized to either receiving advice to discontinue antibiotics if their procalcitonin levels were low or decreasing, or standard of care. And in addition to survival benefit, they also found two reduced days of antibiotics as well as reduced costs associated with the PCT algorithm. However, there was low algorithm adherence. The modified intention-treat analysis showed a fragility index of nine. And mortality benefit was an unexpected finding in this study. And there are some theories for the plausibility of the survival benefit in this study. Authors claim it could be because of antibiotic adequacy. It could be because of timely recognition of alternative diagnosis. It could be because of lower antibiotic toxicity, lowering mortality in the procalcitonin group. Of course, these are theories, and it remains speculative. Technically, we do not need to see survival benefit to use the test. But if we prove in a future trial a plausible mechanism for improvement in mortality, it might result in more goal-directed use of procalcitonin. And potential hypotheses are maybe lower antibiotic duration results in fewer allergic reactions, less drug resistance, less C. diff infection, or it could be an earlier diagnosis of other diseases that may result in a more prompt management of those conditions and better outcomes. The PROGRESS trial, which is a multicenter open-label trial in Greece, tried to answer this very question. This is a recent study in the Blue Journal, 266 patients meeting sepsis 3 criteria. Two-thirds of these were pneumonia sepsis, a third were uroseptic, 16% were bacteremic, and 7% had septic shock. And the intervention was to stop antibiotics if procalcitonin levels were either less than 7.5 or there was an 80% reduction or more compared to baseline. The control group was standard of care. And the primary outcome was infection-related adverse events, which collectively were either a multidrug-resistant infection or a C. diff-associated disease at 180 days. And what they found was there was actually a statistically significant reduction in infection-related adverse events in the PCT group compared to the standard of care group. And in addition, in the secondary outcomes, there were also fewer deaths and lower antibiotic duration of 5 days compared to 10 days. So these findings are quite reassuring and perhaps even inform some of the plausibility around the survival benefit that is being recently reported. However, there are some caveats to consider. The standard of care group received 10 days of antibiotics. Empiric antibiotics were appropriate as high as 98% of the time. The power analysis was based on a probability of MDRO or C. diff infection at 180 days of as high as 30%. And adherence to the PCT protocol was pretty high at 86% among stable patients. And one asks the question, is this sort of idealized setting possible or generalizable across the globe? And maybe some of these elements may have contributed to their findings. Let's take a look. Many parts of the world are moving towards shorter is better. And so on the right-hand side, you see a study by Huang, which is, albeit a study in pneumonia and not sepsis, the antibiotic duration in the control group was only 4 days. And so one wonders whether, is it differences in usual care that in 2022 will dictate whether procalcitonin is in fact useful? Also, in a study that we did for published in Lancet Infectious Disease, where we looked at bacteremic patients, one in five of them was receiving in vitro discordant empiric antibiotic therapy. And more than 50% of this discordance was not due to a resistant phenotype. However, this is very different from the 98% appropriate empiric antibiotic therapy that was observed in that study. And one wonders that in the background of not receiving correct therapy up front, whether shorter courses would produce as good an outcome as was seen in this study. Also, cost-effectiveness is a weak argument to steer ICU's use of procalcitonin. In this modeling study, Schutz and colleagues found that the bulk of cost savings with PCT was actually in the ambulatory and ED setting and really very minimal in the hospital ICU setting. The surviving sepsis campaign 2021 update to the guideline says that for adults in sepsis septic shock, when the optimal duration is unclear, one could use procalcitonin along with clinical evaluation to decide when to discontinue antimicrobials. This is a weak recommendation with low-quality evidence. It's important to note that the bulk of evidence on procalcitonin comes from clinical trials. However, there are some important real-world considerations as well. With growing equivalence between shorter and longer courses based on recent data, a present-day reappraisal of real-world utility of procalcitonin in ICUs is critical. We may want to try combining procalcitonin with antibiotic stewardship programs to decrease antibiotic use, as shown in this single-center study where implementation of procalcitonin into antibiotic stewardship programs resulted in reduced days of sepsis. We all acknowledge that an earlier pathogenic diagnosis is going to be beneficial in sepsis. And there has been a recent emergence of several rapid diagnostic tests and an integration of rapid diagnostics into antibiotic stewardship that has found to result in faster detection, faster escalation, and faster de-escalation. And it might be interesting to see what happens if we actually integrate procalcitonin into this framework. An important piece of guidance for future PCT trials, trialists should focus on one biomarker at a time, rates and rationale for non-adherence in the use of the algorithm should be reported, as well as the involvement of antibiotic stewardship programs and the rapidity of alternative diagnosis due to a negative PCT. And all these factors will increase our ability to attribute the biomarker effect on outcome. This talk would be incomplete without a discussion on procalcitonin and COVID-19. In this pandemic, we have seen high empiric antibiotic use despite low bacterial co-infection burden in hospitalized patients with COVID-19. In fact, there has been a significant variation in the use of initial empiric antibiotic therapy in COVID-19 as seen in this analysis of the PREMIER database. And that's simply because there really isn't any evidence on whether or not initial empiric antibiotic therapy improves outcomes. PCT is elevated in severe COVID-19 as seen here. So it's actually a biomarker of the severity of disease. And there are some possible theories for why PCT levels are high in severe COVID-19. Interferon gamma inhibits IL-1 beta-induced calcitonin mRNA expression, which inhibits procalcitonin secretion. However, chronic ailments like diabetes are associated with a higher risk of severe COVID-19. However, these ailments are also associated with poor interferon gamma responses, which in turn results in a higher procalcitonin concentration. And maybe it's that A results in B and B results in C, and so we are seeing that A results in C. There is expert opinion that procalcitonin could be helpful in limiting the overuse of antibiotics in patients with COVID-19-related pneumonia. But there is unclear utility in ruling out bacterial infection in COVID-19 for the reasons that we just mentioned and additional data are needed. In conclusion, procalcitonin has limited utility in antibiotic initiation and escalation. However, procalcitonin-guided antibiotic discontinuation in critical illness and sepsis at least does not appear to be harmful. We're unable to recommend procalcitonin currently for guiding antibiotic use in COVID-19. And future research is needed on how providers respond to procalcitonin results in the real world. We need to understand better what the generalizability is of the benefits of procalcitonin use in different global regions, because there's a different epidemiology and different pathogen distribution and different treatment patterns in different parts of the world. And we need to better understand the role of procalcitonin in decreasing antibiotic exposure over and above present-day shorter regimens. Thank you very much for your time.

biomarkers

procalcitonin

sepsis

antibiotics

infection

critical illness

Presentation

Select

2022

Advanced

Sepsis Resources