Ladies and gentlemen, please welcome Dr. Timothy Bachman, Editor-in-Chief of Critical Care Medicine, Dr. Howard Bauchner, Editor-in-Chief of JAMA, and Dr. Ahmad Zaheen from the New England Journal of Medicine. ♪ I can hear the whole world singing together ♪ Good afternoon, I'm Tim Bauchman, Editor-in-Chief of Critical Care Medicine. It's my pleasure to introduce the first of this afternoon's light-breaking studies that will truly change our practice. The first study, the Burdens of Sepsis 2012 to 2018, will be presented by Dr. Kimberley Sciretta. This study was released in Critical Care Medicine this past Friday. ♪ Thank you, Dr. Bauchman, and thank you to the Society for giving us the privilege to present our data today. I am Kim Sciretta. I'm the BARDA Solving Sepsis Program Manager. I am a U.S. Federal Government employee. I have no financial disclosures relevant to this presentation, and the disclaimers are the same as those represented in our manuscript. The burden of sepsis is enormous in our country, as the CDC has reported, and the global numbers were also recently revised as published in the Lancet. CDC has determined the at-risk populations for sepsis to include the elderly, those with compromised or weakened immune systems, those with chronic conditions, and children. And we noted that the elderly and the chronic, those with chronic conditions are represented in the Medicare population, so we aim to analyze these beneficiaries. We established a partnership with our sister agency within the Department of Health and Human Services, CMS, and conducted a healthcare quality improvement study. The objectives were to evaluate the burden of sepsis in this Medicare population, estimate the contemporary national sepsis costs, and forecast the future Medicare costs of care for sepsis. The Medicare is a U.S. federal health insurance program. It actually represents greater than 20%, 21% of the U.S. total healthcare costs. It is important to note that the Medicare population, those beneficiaries, are increasing. It increased 22% from 2012 to 2018, which is the interval of our study here, and it's projected to continue to increase to about greater than 80 million in 2030. This analysis leveraged what CMS's data link project, which utilizes pre-adjudicated administrative claims for real-time monitoring and research, and this is of inpatient claims and skilled nursing facility claim data. Our interval period spans seven years from January 2012 to December 2018, and our study is the largest known contemporary data set analyzed of sepsis inpatient admissions with close to nine point, over 9.5 million records analyzed, or claims analyzed. This includes fee-for-service claims data from Medicare Part A and B, and encounter data from Medicare Advantage. For the purposes of this talk and the analysis, we focus specifically on the fee-for-service claims from acute hospital admissions, not the long-term psychiatric or rehabilitation sepsis admissions. It's also important to note that this include both sepsis that was hospital-acquired, so not present on admission, as well as sepsis that was present on admission. We noted that during the interval of our analysis, the U.S. had a coding change from ICD-9 to ICD-10, and so we cross-walked the rates of admission over this timeframe using generalized equivalence mapping. And notice that the crosswalks were similar, so there was no discontinuity. We also were able to stratify the patients by severity using these codings. This research culminated in the publication of three manuscripts released Friday in Critical Care Medicine. This talk will primarily focus on paper number one and paper number three. Initially, we explore the incidence of sepsis in this Medicare population, and saw that the sepsis claims increased steadily over this time period, increased almost 40% during this rate of 40% per capita of the Medicare beneficiaries. This is in comparison to the Medicare population increasing only 22%. We also noticed across all severities, there was seasonal variation that appeared during the winter months that appeared to correlate or could correlate with respiratory illness during that time. The chart on the right graphs the counts of sepsis per month at these severities. In the next slide, I will show the totals per month of sepsis, and you can see that the counts also increased over this time period from 770,000 in 2012 to over 1.1 million in 2018, and that is graphed here on the right-hand side. So, sepsis is increasing across all severities, suggesting that a sepsis is more common. However, there has been questioning as to whether this could be due to overcoating. So, we explore the proportion of sepsis across these severities and noted that the proportions of septic shock and severe sepsis remain stable. You might expect a disproportionate fall in these proportions of septic shock or severe sepsis and an increase in the non-severe sepsis counts, but we don't see that. So, therefore, we don't think overcoating is a major contributor. We then looked at the proportion of admissions that were arising in the hospital versus being present on admission, and we saw that the rise seen in incidence appears to be exclusively due to the present on admission group. This fraction increased from 87% in 2012 to 93% in 2018, and this is noted in the red on the figure on the right. So, sepsis continues to rise in the community. Hospital-acquired sepsis, on the other hand, so not present on admission, decreased during this time period from 13% to 7.5%, suggesting that hospital practices are indeed improving. We next looked at mortality trends, which seemed to correlate with sepsis severity. We looked at mortality during or within the one week of discharge of a hospital, six months, and one in three years. And the good news is that mortality is decreasing over the seven-year period, suggesting sepsis diagnosis and treatment as effective. But what was astounding is there's a very high three-year mortality. For septic shock, this is 75%. For non-severe sepsis, this is 60%, and this compares to only 40% in hospital stays that are not sepsis-related. Lastly, we wanted to explore the contemporary cost of sepsis in our nation. The last time this was reported was in 2016 based on 2013 analysis of HCUP data. The cost of sepsis at that time was $23.7 billion, and the Medicare population was estimated to be 61.5% of those costs. So initially, we looked at the average paid claim per admission. So for inpatient costs, this claim payment decreased over the time interval. However, contrary to this, the SNF costs rose slightly over this time period. So if we total the costs for sepsis in this Medicare population in 2018, utilizing the claims that we know for fee-for-service claims in blue and the imputed costs for Medicare Advantage in green, we have inpatient costs on the left, SNF costs on the right, and this counts to or accumulates to $41.5 billion just for 2018. We now wanted to project into the future into what the cost could be for 2019. So utilizing regression models, we were able to project forward, again, inpatient on the left, SNF costs on the right. We're now at $44 billion. But this is just Medicare costs. This does not include costs for the entire nation. So we sought to approximate this. We know from the 2016 study that the Medicare population was estimated to be about 61% of the total costs. We estimated the 2009 Medicare inpatient claims to be $35 billion. So we thought these were approximately about two-thirds of the costs for Medicare. We used a more conservative number here to account for the aging and the growth of the U.S. population. So from this, we can calculate the total inpatient estimated cost in the nation to be $53 billion. But this does not account for the SNF costs. So if we add the SNF costs calculated from or estimated from 2019, which is $9 billion, we are at $62 billion. Now, note this is a conservative rough order of magnitude lower bound estimate. And the reason why we say this is because this doesn't cover all of the costs of sepsis. It doesn't cover those initial pre-hospital costs. It doesn't cover the physician fees. It doesn't cover the costs occurring in the VA or in the DHA. It doesn't cover SNF costs in the private sector. And it doesn't cover costs post-discharge for home care. So the cost of sepsis could well exceed the $62 billion. So in conclusion, we are seeing that sepsis admissions are rising within the Medicare population, almost double the rate of the Medicare beneficiary population growth. We are seeing a decrease in mortality, which is good news. But the sepsis-associated mortality at three years is still astounding. We're at 60 to 75%. The cost per admission is also decreasing. But if we total the total cost of sepsis in our country, we're at the $62 billion. We knew there would be interest in expanding the analysis to other datasets. So we developed a framework to standardize this process and made this openly accessible. This is available as supplement one in paper number three. To learn more about the trajectories of sepsis patients from paper number two, my colleague Dr. Simpson will be presenting in tomorrow's late-breaker abstract session. If you want to learn more about these studies, you can contact my colleague sitting here to the left of me on stage. And he can provide more information on these publications. I want to acknowledge the BARDA Drive Solving Sepsis Team. You can contact us at solvingsepsis at hhs.gov or at drive.hhs.gov slash solvingsepsis. Thank you to all the authors on this paper from BARDA, ASPR, CMS, and Acumen. This is how you can contact BARDA. Thank you very much. Thank you, Dr. Sciretta. Our next speaker this afternoon is Dr. Jerry Zimmerman, who will be presenting Outcomes After Community-Acquired Pediatric Septic Shock. This study was also released this past Friday in Critical Care Medicine. Dr. Zimmerman. Well, good afternoon, everyone. And thank you to Critical Care Medicine and the Society of Critical Care Medicine for allowing me to present the results of the Life After Pediatric Sepsis Evaluation Study. I'm actually going to talk about two studies. The first related to the trajectory of functional status and quality of life following an episode of community-acquired septic shock. And then secondly, the variables in the ICU that predict these adverse outcomes. I would be remiss if I did not thank the National Institutes of Child Health and Development for funding this study. Well, we all know that sepsis is a big deal in the world in which we live. Every three seconds, someone in the world dies of sepsis. And it turns out that about 40% of these deaths are children less than five years of age. This is a common cause of death in hospitalized patients. And as mortality from sepsis has decreased over the last several decades, particularly for children, researchers and clinicians are increasingly interested in long-term morbidity among children who survive sepsis. SOLAPS was a prospective descriptive cohort investigation that involved 12 quaternary PICUs in the United States and enrolled patients between 2014 and 2017. The subjects for this investigation were children with septic shock requiring vasoactive inotropic support, illness severity, demographic, organ dysfunction, and resource utilization were collected during ICU stay. And then functional status as well as health-related quality of life were assessed sequentially during the first year following admission for the sepsis event. These are the performance sites which leads me to tell you or thank the patients and families particularly, the research coordinator infantry and the participating principal site investigators that made this investigation possible. The strength of this investigation is that all patients were assessed for their baseline status and all subsequent measurements were really judged against that. So every person, every subject served at their own control. After this initial assessment, patients were assessed at seven days, one month, three months, six, and 12 months after the sepsis event. The expected enrollment is shown on the red line and the actual enrollment in the blue line underneath. Enrollment into the study ceased as of June 2017. And for this study, 832 children were screened. 632 were eligible and approached. Of these, 90% were approached and 69 in the end, 69% of these families agreed to participate in the study. Hospital, in-hospital mortality was about 9%. So here's the infectious disease status of this cohort. About half had bacterial infections, 45% had viral infections, 20% had a combined bacterial and viral infection, and then 30% of this cohort had neither a positive culture nor positive polymerase chain reaction for a virus. This study enrolled basically the spectrum of age among children. 45% of the subjects were girls and 17% were immunocompromised. Here's some information about illness severity, organ dysfunction, and resource utilization. These children required as a median three and eight days of vasoactive inotropic support and mechanical ventilation, nine and 16 days of PICU stay and hospital stay. Thirty-eight and 27 children required renal replacement therapy and extracorporeal life support, each for a median of eight days. So the first specific aim of this study was to track the trajectory, how these children turned out after their sepsis event with respect to their functional status and health-related quality of life. We first ascertained a measure of gross functional status utilizing the pediatric, cerebral, and overall performance category score, directing your attention to the right. Only 38% of this cohort had a normal functional status at baseline, and this correlates with the fact that over half had a significant chronic complex comorbid conditions. This functional status plummets, as you would expect on day seven when most of these kids were in the ICU. But even a month later, the majority of these children, by either measure, have an abnormal functional status. We also looked at functional status with a more granular measure, the functional status scale. You see here that at baseline utilizing this tool, about half of these kids were normal. Again, the status deteriorated with critical illness. But again, at one month following admission to the PICU for this sepsis event, 65% of these children demonstrated an abnormal functional status. Two measures were utilized to assess health-related quality of life. The first, the pediatric quality of life inventory, well-validated tool. The second tool that was used is to Stein Jessup functional status scale, a measure of general health. And this tool was chosen by families who thought that perhaps some questions on the PEDS QL just did not make sense for their children who were typically severely developmentally delayed. These two figures show the raw scores for these two instruments over time. On the left, you'll see the PEDS QL. And I might point out that at three months, as a median of the total score, the group was sort of back to their baseline status, on average. And for the children assessed with the FS2R, these children, as a median score for the entire group, was back to baseline by approximately one month. The other way of looking at this, which is maybe more clinically easy to understand, is looking at the proportion of those children who do not return to their baseline health-related quality of life. So here's the data for PEDS QL. At day seven, when these kids are in the ICU, 70% of these kids are below their baseline quality of life. And then you can see how this tracks out over time. At one month, it's 60%. And at 12 months, it's still 40% of these children who have not yet achieved their baseline health-related quality of life. Similarly, for children with severe developmental delay assessed with the alternative score, at day seven, about 60% below their baseline. And then tracking their progress over the ensuing year, so that at 12 months in this cohort, 30% are still below baseline. You may have also noticed that there is not much happening between three months and 12 months. Overall, if you look at mortality, the cumulative mortality over the one-year period was 13% for this cohort. And the failure to return to baseline health-related quality of life was 35%. Thirty-five percent of these kids had not yet achieved baseline quality of life. The second part of this was to examine PICU risk factors for these adverse outcomes. It's important to define what we're talking about, death as all-cause mortality, persistent serious deterioration of quality of life, 25% below baseline, and poor outcome as either one of these. At one month, 8% of these children had died, and 29% had severe quality of life deterioration for poor outcome of 37%. And at three months, more deaths, fewer quality of life, severe deterioration, but still 28% with a poor outcome. We looked at a variety of ICU variables that might affect or influence these adverse outcomes, and these consist of illness severity, organ dysfunction measures, and resource utilization. There are multiple models that we derive, but the one that we finally settled on was this one. The first variable, summation of PILOD scores, which generally would quantitate the intensity and severity of organ dysfunction. The second, the highest vasoactive inotropic score, which would reflect the highest severity of septic shock, and lastly, as you'd expect, any acute pathologic neurologic event. These three factors were all independently associated with poor outcome. So, the majority of children with contemporary septic shock present with chronic comorbid conditions. These children require about three days and eight days of vasoactive support and mechanical ventilation, end up staying in the ICU and hospital for nine and 16 days, respectively. Mortality in a hospital is 9%, and at one year, 13%. The take-home message, the big one here, is 35% of these children after one year following the septic shock event still have not achieved their baseline health-related quality of life. Poor outcome was noted in 37 and 28% of these kids at one and three months after the septic shock admission. And the variables that are associated with these adverse outcomes are measures of individual and cumulative organ dysfunction. So in the ICU, we treat organ dysfunction, and we feel pretty good when we save a kid from septic shock. But survival no longer expresses the complete impact of septic shock on these children. For children ultimately to survive sepsis, clinicians and investigators alike need to focus on interventions that not only improve survival, but facilitate long-term optimal health-related quality of life. Thank you very much for your attention. You have to go up there. Thank you. I'm Howard Bochner, Editor-in-Chief of JAMA. And I'm pleased to introduce the next speaker, Dr. Bob Lee, who will be presenting a paper that will be published simultaneously in JAMA, entitled Association of Physician Orders for Life-Sustaining Treatment with ICU Admission Among Patients Hospitalized Near the End of Life. And just one comment about the paper. It was submitted almost a year ago, but the data only went through 2015. And we're very acutely aware that many issues around death and dying, both in the ICU and American hospitals, have changed over the last couple of years. So I just want to thank Bob and his co-investigators for updating the data. So it's really quite current. It goes now through 2017. Thanks very much. Great. Thank you all so much. So my name is Bob Lee. I'm an instructor and palliative care researcher at the University of Washington. And I'm honored to present to you our recent work. And thank you very much to the Society and also to the editors of JAMA for this opportunity. I have no conflicts of interest to disclose, although I always like to take this opportunity to acknowledge our funding sources. So by way of background, many patients with chronic life-limiting illness don't necessarily want all of the available intensive care that we have to offer. And in the United States, patients use a variety of measures to document their wishes for treatment limitations. One popular paradigm that has been implemented throughout the country is called the PULST. PULST, for those of you who are not familiar, stands for Physician Orders for Life-Sustaining Treatment. And it's similar to an advanced directive, but it differs in that it contains some discrete orders that are actionable by healthcare providers anywhere. They've been implemented throughout the United States with varying degrees of uptake. And since we're a country that can never agree on anything, they go by a number of names, PULST, MOLST, COLST, et cetera. This is the Washington State PULST here. And I just want to highlight two of the key orders that are on it. First of all, there's an order for whether or not to give CPR in the event of cardiac arrest. And then there's an order for medical interventions. This is an order for what type of medical care the patient would want in any illness state other than cardiac arrest. And their choices range from comfort measures only, which is essentially hospice care, only symptom management, to limited interventions, which basically describes avoidance of intensive care, to full treatment or full-intensity care. And these treatment-limiting PULST orders have been associated with less in-hospital death and less CPR, but we don't know as much about what happens with patients who have PULST forms after they arrive at the hospital. And so we set out to address two research questions. The first one is, for patients with PULST who are hospitalized near the end of life, are their PULST orders actually associated with the intensity of care that they receive? And the second question that we looked at is how often do patients receive intensive care that's actually discordant with their PULST? And what are the risk factors for receiving that type of discordant care? So to address these research questions, we conducted a retrospective cohort study of decedents. All of them had chronic life-limiting illness, and we defined this using the Dartmouth Atlas conditions. This is a list of nine diagnoses that are associated with about 90% of the deaths in the Medicare population. They all had to have been hospitalized within the last six months of life at one of our study hospitals at UW Medicine. And they had to have completed a PULST prior to their last hospitalization of life. The goal here was really to identify patients who had a PULST that was timed so that it would inform the care they received during that last hospitalization of life. And they had to have died between 2010 and 2017, and we measured this using death certificates. Our exposure was the PULST order for medical interventions that we talked about earlier. So comfort measures only versus limited interventions versus full treatment. And we also looked at a number of potential confounders. There's a few demographic variables, and then namely a few things that I want to point out. The amount of time that passed between when they completed the PULST and when they were admitted to the hospital. The person who signed the PULST, whether it was signed by the patient or whether it was signed by a decisional surrogate. And a history of dementia, history of cancer, and also whether or not the patient was admitted for a traumatic injury. Our primary outcome was ICU admission during the last hospitalization of life. And we examined that last hospitalization with the intention that this would capture the patient's terminal illness episode. And recognizing that not all ICU admissions are alike, we also looked at the secondary outcome of life-sustaining treatments. Mechanical ventilation, vasopressors, new dialysis, or CPR. And we measured these outcomes from the electronic health record. We used modified poisson regression with robust error variance to estimate relative risk. And we had a small percentage of missing data in some of our covariates, and so we used multiple imputation to impute those. So we had 1,800 participants, of whom 400 had pulse orders for comfort measures only. And 760 had pulse orders for limited interventions. And I point those two groups out because those patients are the patients who have pulse forms that specifically say that they should avoid intensive care and we'll refer to those people as treatment-limiting pulses. Here's some of their characteristics. So this is an older cohort, as would be expected for a patient, for a cohort of people with chronic illness. Mean age of 70, about 40% female. 26% were of minority race or ethnicity. And this is relatively multi-morbid cohort. The median number of Dartmouth-Atlas conditions they had was three. So this is an analysis that's looking at the association between the primary outcome of ICU admission and the predictor of pulse order. The relative risks that are shown here compare patients with these treatment-limiting pulse orders to patients who had full treatment orders. And we found that treatment-limiting orders are, in fact, associated with less intensive care. Those who selected limited or comfort measures only were less likely to be admitted to the ICU than patients who selected full treatment. And similarly, when looking at the secondary outcome of life-sustaining treatments, those with limited or comfort-only pulse were also less likely to receive these life-sustaining treatments. And taken together, I think this suggests that pulse orders are associated with the likelihood of being admitted to the ICU or receiving life-sustaining treatments near the end of life. So that's great. It's nice to know the world works. But we also noticed that despite these relative risks, there's about a third of patients in the comfort measures only group who are still being admitted to the ICU near the end of life. And 46% of patients in the limited interventions group. And I'll remind you again, these are patients who checked off a box saying that they wanted to avoid intensive care. And some of these admissions might have been less intense than others, but about 15 to 20% of these admissions involve mechanical ventilation, vasopressors, CPR, or dialysis, mu-dialysis. So there's a small but significant percentage of patients who are receiving life-sustaining treatments that are discordant with their pulse orders. And when we think about what that all means, it's important to recognize that pulse-discordant care isn't necessarily always inappropriate or wrong. You know, patients, our ultimate goal is to provide patients with care that's aligned with their goals and values. That's really hard to measure in a research setting. But our study is examining pulse-discordant care, which is much easier to measure, but may or may not be appropriate, depending on the circumstances that led up to it. And so we were interested in exploring why patients might receive pulse-discordant care. And so we did some analyses looking at the outcome of pulse-discordant care and looking at potential risk factors for this outcome. These results are for patients who had pulse orders for comfort measures only. And highlighted in yellow, you see that patients with cancer or dementia are significantly less likely to receive pulse-discordant care, whereas patients admitted for a traumatic injury highlighted in red were more likely to receive pulse-discordant intensive care. And similarly, in the analysis of risk factors for the outcome of pulse-discordant intensive care among those with limited interventions orders, we see, again, that patients with cancer are less likely to receive intensive care and that those admitted for trauma are more likely to receive pulse-discordant care. We also found a small but significant association between intensive care and younger age. So putting those last two slides on one slide and just kind of thinking about, well, what does all this mean? One of the big challenges with pulse, and really with any paradigm in which we're making medical decisions in advance, is that those decisions have to be made with hypothetical context. In other words, they have to be made out of context. And those hypotheticals can be difficult to pin down for certain patients. Patients with cancer or dementia may have more predictable illness trajectories near the end of life, and compared to, say, a patient with CHF or end-stage renal disease, it may be easier to allow to do care planning for these patients in more concrete terms. Whereas on the opposite side of that spectrum of traumatic injury is something that's actually quite common amongst chronically ill patients, but nobody envisions it, nobody talks about it, nobody expects it. And we think that this role of predictability may play an important role in pulse-discordant care. Our study has a few key limitations. First of all, we only examine patients who are hospitalized in the last six months of life. A substantial percentage of patients we believe are never hospitalized near the end of life, particularly those with comfort-only pulse. And so we are likely to overestimate the incidence of pulse-discordant care amongst all pulse users. We only examine care that was delivered at our two study hospitals. And so it's possible for somebody to have received intensive care from another entity that we did not capture. This would have led to underestimation of the prevalence of pulse-discordant care. We did do a sensitivity analysis where we restricted to patients who died in our hospital, and that analysis did corroborate our findings. And lastly, Washington's a state that has a pretty mature pulse implementation. It's an intervention that has high uptake there, but our findings may not generalize to other health systems or to other settings. So in conclusion, we found that for patients with pulse who are hospitalized near the end of life, treatment-limiting orders were associated with significantly lower rates of ICU admission. But 38% of patients with treatment-limiting pulse nevertheless received pulse-discordant intensive care. And so I think our take-home message from this is that pulses are probably helpful for patients with chronic illness, but there's also room for improvement. For who do we complete pulse? How do we complete pulse? How do we communicate them across the healthcare system? And for those patients who are having unanticipated illnesses, unanticipated crises, how do we help patients and families make the best decisions in the moment? And those are some opportunities for future work. I want to thank Dr. Cross and Dr. Curtis, who are sitting in the third row over there. They are my primary research mentors and senior authors on the paper, as well as the rest of the co-authors from UW and from OHSU. Thank you very much. Appreciate it. Thank you, Dr. Lee. Hello, everyone. I am Dr. Ahmad Zaheen, representing the New England Journal of Medicine. I'm pleased to introduce our final speaker this afternoon, Dr. Pele Toft, who will be presenting trial of targeted non-sedation versus sedation in critically ill ventilated patients. The study is being simultaneously released in the New England Journal of Medicine. Thank you for inviting me to this meeting and giving me the opportunity to present this randomized controlled trial for which I am the sponsor. We compared non-sedation to light sedation in 700 patients. We compared non-sedation not even to light sedation, but to light sedation with a daily wake-up trial, during which the patients were weaned intensively from the ventilator. In other words, we compared with the best of the best. This slide, you can see the non-sedated patient. He's mobilized to his chair, and he has his wife to support him. And on the right side, you can see the hand of the nurse. This is important because most of our units will have a nurse-to-patient ratio of one to one. And the nurse is always bedside. She's not in a remote area. She's bedside. You can only do non-sedation if it's a cooperation between the nurses and the doctors. And I have no conflict of interest because it's not a drug I'm trying to sell. It's a change of culture. After our opinion, an oral tube, the modern oral tubes and modern ventilators, it's not an absolute indication for sedation. And why did we perform this multicenter study? Well, 10 years ago, we performed a single-center study where we can show that patients who are non-sedated were shorter time on the ventilator, shorter time in the ICU, and shorter time in the hospital in a single center. In this single-center study, we also found, to our surprise maybe, a nearly significant advantage for survival with non-sedation. This was hypothesis generating. If this was true, we would need 700 patients in a multicenter study to show that non-sedation would improve survival. So the primary outcome was 90 days mortality. Secondary outcomes were thromboembolic events, acute kidney injury, coma and delirium-free days, and others. It was a Scandinavian study, only eight ICUs in Denmark, Norway, and Sweden, which used only eight centers because it was a change of culture, and it's more difficult than giving a drug. So with eight centers, we thought we could control what was going on. At baseline, the patients were well-characterized, except for a significant higher PET-C2 score in the non-sedated groups. In other words, they were slightly more ill. What did we find? Well, concerning mortality, there was no significant difference between the two groups. Maybe it's easier to show them a figure. There's no difference in mortality, so we were not able to confirm our hypothesis that non-sedation was superior concerning survival. But what about the secondary outcomes? Well, thromboembolic events, as defined as deep vein thrombosis and pulmonary embolism, there's actually only one case in the non-sedated group compared to 10 in the sedated group, which was significant. It's a rather low figure, probably because all the patients received low molecular weight heparin. A number of ventilator-free days. To our surprise, there was no difference between the two groups. A number of ICU and hospital-free days, nearly identical ICU-free days. And concerning hospital-free days, there was no significant difference. But coma and delirium-free days. During the stay in the ICU, there were three coma and delirium-free days compared to one in the sedation group, which was significant. Even if we extended the period to 28 days, there was still significantly more coma and delirium-free days in the non-sedated group. Non-sedation and the kidney function. In our single-center study, we were able to show a better RIFLE score, a lower RIFLE score in the non-sedated group. So a priori defined goal was the highest RIFLE score in the two groups. As you can see, there was no difference. We measured other things concerning the kidney. This was not a priori defined, so we have to interpret the results with caution. But there was significantly more days with maximum highest RIFLE score in the sedated group. The sedated groups, there were more patients on CRRT. And when they were discharged from the ICU, a higher percentage of patients in the sedated group needed hemodialysis. Physical function. We did a priori defined subgroup analysis of 200 patients. And as you can see, the physical strength, the hand grip strength, was superior in the non-sedated group when they were extubated. And it was superior when they were discharged from the ICU. The Basel index is a measurement of physical independence. The higher, the better. And you can see that it was significantly higher in the non-sedated group. In other words, the non-sedated patient has a higher degree of physical independence when they left the ICU. How did we perform? This is our single center study. And you will see that the non-sedated group had a RAS score between zero and minus one. In other words, they're not completely awake. They may be sleeping. But you can easily wake them up. What about the multi-center study? Well, we didn't perform as well, especially on day one and maybe on day two. The RAS score was lower, indicating the more patients were sedated, especially on the first day in the non-sedated group. So in this multi-center study, during the first day, 27% of the patients were sedated. And overall, 38% of the non-sedated patients were sedated at some point during their stay in the ICU, most often for only one day. And this is more than our single center study, where only 18% received sedation in a non-sedated group. That might explain why we couldn't find a reduced stay on the ventilator. We compared with the best of the best. And the control group was arousable sedation. And in the control group, we performed a daily wake-up trial, where we weaned the patient intensively. Well, there must be some adverse events. Yes, there is. We defined serious adverse events as accidental extubation, which required intubation within one hour. There were four patients in the non-sedated group and one in the sedated group. Total number of accidental extubations was 8.9% in the non-sedated group versus 4% in the sedated group. But it's not zero in the non-sedated group. It still happens. Well, if there are some adverse events and we don't prolong or improve their survival, why should we use non-sedation? Well, it's another form of care of the patients, where patients are able to communicate with the staff. They might be able to enjoy food. In the evening, they can look at the television instead of being sedated. And they can be mobilized to do other activities. They are mobilized, and they can write their opinion about the treatment to a doctor. And this way, you have a two-way communication. My conclusion is that light sedation with a daily wake-up trial was comparable to non-sedation as regards to mortality. But non-sedation reduced the number of tromboembolic events. Non-sedation might, I say might, improve kidney function. Non-sedation increased coma and delirium-free days. And non-sedation preserved the physical function to a larger extent. So I think my main message is that we have to individualize patient treatment. Many patients would benefit from non-sedation, and some would benefit from light sedation with a daily wake-up trial. We have to respect patient autonomy and try to establish a two-way communication with our patients in 2020. Thank you for staying awake.

Dr. Pele Toft

randomized controlled trial

non-sedation

light sedation

patient outcomes

90-day mortality

thromboembolic events

physical function

Sepsis

Resuscitation

Shock

Advance Directives

Presentation

Select

Professional

Associate

2020

Intermediate

Advanced

Sepsis Resources