Well, good afternoon, and thank you all for coming in on what seems like a very busy Monday afternoon, evening for SCCM activities. I'm going to start by saying that steroids and sepsis is interesting, and not only is it interesting, it's one of those things. When I talk to my trainees, or I talk about septic shock, and I bring up steroids, I always start by saying 50 years ago we were talking about steroids and sepsis. 50 years from now, or maybe even after I'm no longer in this world, we'll still be talking about steroids and sepsis, and I say mostly that it'll still be an open debate. If you look at the history of steroids and sepsis, we have gone from all kinds of doses, all kinds of agents, to lower doses, appropriate time of usage, and if you still survey practitioners across the world, what you and I do, and how we handle steroids in a patient in septic shock will be different. In fact, if I ask the audience right now, when do you start your hydrocortisone? At what point do you taper it? At what point do you start it? I will get a variety of answers from everyone. I can almost guarantee that. To that point, it is my honor and pleasure to introduce you all to the man who has probably lived with steroids for a large chunk of his lifetime. It is an honor and a pleasure to introduce Dr. Bala Subramanian Venkatesh, or Bala as he likes to go by, and I have to tell you, we met first in Belfast, Ireland, where I was presenting our work on angiotensin 2, and he, at that meeting, I had the honor of sitting in the audience as he broke the results of the adrenal trial to the world. I have to say, it was interesting. It was January 2018 when this happened, and I'm not going to tell you what the results showed. You guys probably know that, but then two months later comes out another trial that showed something slightly different, and people were again thinking, well, what do we do now, right? And I guess, you know, I'm sure all of you have all kinds of questions for him, so I will very quickly let him get started, but he insists I don't do a long introduction for him. I'm still going to do a little bit of introduction because I insist on telling you guys all about this man. He is a professor of intensive care at the University of Queensland and a preeminent specialist in intensive care medicine at the Princess Alexandra Hospital and deputy director of the intensive care medicine system at the Wesley Hospital. He's also president of the College of Intensive Care Medicine, and obviously, I told you this, he's the principal investigator of the funded adrenal trial, which is a multi-center trial looking at outcomes of steroid septic shock, and he is a mentor, a friend, and a personal source of inspiration for me, and I'm guaranteeing you guys that when you leave the audience at the hall today, you will probably have a lot of good points to take home and hopefully points that will improve your clinical practices of how you handle steroids and septic shock. His title says, is it time to end the debate on steroids and septic shock? I probably feel no, but let's see what Bala has to tell us. Welcome. Thank you. Thank you, Aseesh, for that very kind introduction, and it's a great honor to be here, and this is, as Aseesh mentioned, this is a very hotly debated topic, the role of steroids and septic shock, and so what I plan to do, and I think it's important to look at this question in detail. I guess the key question in people's minds is, have the two recent trials clarified the role or confused the picture further? And one of the key differences between the two recent trials, the adrenal and the aprox, was the use of fluid recorder zones. I'll talk to you about that. The third key question is, what should be the trigger for initiating corticosteroids, and are there any specific pressors which are more responsive to steroid usage as compared to others? Should we taper corticosteroids? Is there any need to do a corticotropin test before you initiate steroids? Is there a role for repeat dose of steroids if someone had a lapse of shock? And I guess I also just want to let you know what we have done in the last two years since the publication of adrenal, what additional work has been done, and what the agenda for future research should be. Now, these are my disclosures. I have grants from the Medical Research Council of Australia, and I have a practitioner fellowship, and the adrenal trial was funded by the NHMRC, but I have no other conflicts of interest. So, I know it's a one-hour session, but I'm not going to speak for one hour. I'll speak for about 35, 40 minutes. I'll have 15, 20 minutes for you to ask questions. So, the story began in 1923. This was a very clever study by Scott, and essentially, they took rats, and there are two groups. One group were adrenalectomized. The other group were controls, and bacterial injections were performed to both groups of animals, and the group which had adrenalectomy died, uniformly died, 100 percent fatality. The other group survived, and therefore, the notion that steroids might be important for the stress response of any illness began to appear, and then in 1940, we had the famous work by Heinz Selle, who is the father of stress physiology, and where he reported clearly what the adrenal glands did during stress, and the first clinical trial of steroids happened in 1976, Schumer's paper, but the sort of two landmark RCTs happened in the 80s, and these trials looked at high-dose corticosteroids and septic shock. The first was conducted by Charlie Sprung, and the second by Roger Bohm, and they used 30 milligrams per kilo per day of methylprednisolone, which is equivalent to using 10,000 milligrams per day of hydrocortisone. Essentially, what they showed was that the use of high-dose steroids, which was being practiced until these trials came out, the use of high-dose steroids were associated with either no benefit or evidence of harm in terms of a trend towards increased mortality or adverse effects, so following these two trials, high-dose steroids were clearly dead in the water. During the 90s, there was a sort of a reduction in the enthusiasm for using steroids until two papers appeared, which suggested that lower-dose steroids, not high doses, but super physiological dose of steroids, something like 200, 300 milligrams per day might be beneficial, and the drivers for these two papers, for this work, for the research in lower-dose steroids, was one, was the finding that there was improved pressure responsiveness amongst patients to catecholamines when they were given steroids, and secondly, there was a suggestion that there could be relative adrenal insufficiency in patients with septic shock, and therefore supplementing with corticosteroids would be beneficial. So these two papers provided sufficient justification for the conduct of two more trials, randomized controlled trials. These used low-dose steroids, and everybody is probably familiar with the ANAND trial, which was published from France, published in 2002 in the JAMA, and they used hydrocortisone plus flutricortisone in one arm versus placebo in the other arm, and the second trial was Charlie Spang's trial, the Corticus trial, published in 2008, which just looked at hydrocortisone alone versus placebo in patients with septic shock. The results of these two trials debated substantial, generated substantial debate. The ANAND trial basically showed that in the intention to treat analysis, there was no differences, there were no differences in mortality between the steroids and the placebo arm, but in the subgroup of patients who were deemed corticotropin non-responders, or those who were deemed to be adrenally insufficient, there was a survival benefit with the use of steroids. On the other hand, the Corticus trial showed no differences between the two groups on the intention to treat, nor in the subgroup of patients with corticotropin non-responders. So people looked at these two papers in great detail. You can see the sample sizes there, 300 versus 500. The medical cohort was much higher in the ANAND group, as opposed to two-thirds medical, one-third surgical. It was the converse in the Corticus. The treatment, again, ANAND used fruticortisone, Corticus only used hydrocortisone, and what's important is that there was a atomic usage in about a quarter of the patients in both trials. So there were some question marks with the power. Did the trials have sufficient power to detect a mortality difference? And particularly, the Corticus trial was stopped prematurely because of lack of equipoise to enroll patients. The second problem, as I said, was the divergent results. The third was a quarter of the patients received atorvidate, which is a known adrenal suppressant. And interestingly, the ANAND trial reported no adverse effects, but the Corticus reported an increase in adverse effects in the intervention group. So following these two trials, the people were more confused, and so there was what's called a global assessment of the utilization of low-dose steroids, and that was published in what's called a progress registry. Nearly 12,500 patients with septic shock were evaluated, and what they showed was there was substantial variability in practice and marked clinical uncertainty. And so this is just to give you the usage of low-dose corticosteroids and septic shock in different parts of the world. It was as low as 25 percent in the oceanic region, as opposed to 60 to 70 percent in Latin America. So the time was really ripe. This question had to be addressed, and it had to be done where you had sufficient sample size, you had clinical equipoise, as in if you enrolled a patient into a clinical trial with septic shock, the clinician was prepared to withhold corticosteroids as part of routine care, and it also had to be done in an atorvidate-free environment. So in Australia and New Zealand, A, we don't have atorvidate. It's not available. B, we performed a survey, and there was clinical equipoise. Clinicians were prepared to enroll patients into these trials, and C, we were used to doing large clinical trials. So this trial was funded by the NHMRC, the Health Research Council of New Zealand, and the National Institute for Health Research in the U.K., and this was published in the New England Journal of Medicine in March 2018. Now, we enrolled 3,800 patients with septic shock, so as compared to 300 versus 500, just to give you an idea. So I'll just go to the results straight away, and basically, just to give you what we did. So the inclusion criteria were mechanically ventilated patients with septic shock and no atorvidate use, and we had a standard exclusion criteria for sepsis studies, the same ones we used here, and we gave corticosteroids for up to seven days or until discharge from intensive care, whichever was earlier. Now, the primary outcome was day 90 mortality, and it was 27.9 percent in the hydrocortisone arm, 28.8 in the placebo arm, and you see the confidence limits there and the p-value, which is not significant, and when we adjusted for stratification variables and additional covariates, again, the statistical significance did not change. We then looked at the time to death, okay? We know the number of people dying between the two groups is not different, but do they die faster in one group versus the other? So we looked at the Kaplan-Meier curves, and again, there was no difference between the two groups. We then looked at the primary outcome in six pre-specified subgroups. One was the sex, the admission source, the severity of shock as determined by the catecholamine source, and the site of sepsis, because pulmonary sepsis is thought to respond better to corticosteroids, and severity of illness, Apache score greater or less than 25, and the time from onset of shock to randomization, and again, we did not see any differences across the six pre-specified subgroups in terms of a treatment effect. The trial was conducted in five countries, Australia, New Zealand, Denmark, Saudi Arabia, and the UK, and we looked at the primary outcome by region, and again, the polynomial p-value was not statistically significant. When we look at the secondary outcomes, they told a different story. Hydrocortisone was associated with a more rapid resolution of shock by one day, a shorter duration of the initial episode of mechanical ventilation, an earlier time to ICU discharge, and a reduced frequency of blood transfusion, and all these were significant after adjustment for multiplicity. There were no differences between the two groups with respect to recurrence of shock, need for renal replacement, recurrence of mechanical ventilation, and superinfection rates as defined by nuance and bacteremia, or fungenia. There was a small but a slightly higher incidence of adverse effects in the hydrocortisone group, but these were mainly metabolic in nature and did not impact on patient-centered outcomes. So the principal findings in adrenal were that hydrocortisone did not lead to a significant reduction in mortality at 90 days. The effect on mortality did not differ in any of the six pre-specified subgroups. Hydrocortisone was associated with a more rapid resolution of shock, shorter duration of the initial episode of mechanical ventilation, earlier time to ICU discharge, and a reduced frequency of blood transfusion. We then reported the long-term outcomes of the adrenal trial, the six-month mortality as well, in the New England Journal, and it's again not different between the two groups, 31.5 versus 31.8%, not statistically significant. In the same issue of the journal, the APROX trial was reported. The APROX trial was a, I'll come to the trial design in more detail later on, but essentially an RCT which compared the combination of hydrocortisone plus glutrocortisone versus placebo in patients with septic shock, and they reported an improvement in 90-day mortality, a six-percentage point difference, absolute risk reduction, 43 versus 49%, and they also reported improvement in secondary outcomes in the steroid group, very similar to what we observed in adrenal. And after these, we then there was a, we published a systematic review and a meta-analysis, again, which did not show any, actually did not show a mortality benefit for the use of steroids, and, but certainly evidence in terms of mechanical ventilation and shock reversal. So the question now is, and this is the thing which has been the matter for debate, why the adrenal did not show a difference in mortality and why the APROX did. And we summarized these arguments in an issue of critical care medicine last year. So I think it's important to look at the differences between the two trials. The adrenal was a, I think the trial design, the adrenal was a parallel group RCT, whilst APROX was a two-by-two factorial design. It started off as a two-by-two factorial design, looking at activated protein C, hydrocortisone plus glutrocortisone, the combination of the two, and placebo. Halfway through the trial, activated protein C was withdrawn from the market, and it then got converted to a parallel group RCT. The second main difference was mechanical ventilation was an inclusion criterion adrenal, not in APROX. A minimum dose of preso was specified in APROX, 0.25 mics per kilopermit of noradrenaline, not so in adrenal. Etomidate was an exclusion criterion in adrenal, not so in APROX. The proportion of medical patients with pulmonary sepsis was significantly different between the two, 35.1% in adrenal, 59% in APROX. And proportion of patients receiving vasopressin was again different, 16.2% in adrenal and 0.08% in APROX. And finally, of course, the interventions, glutrocortisone was used in adrenal, not so, sorry, in APROX, not in adrenal. So the first question, and it's been argued that the APROX showed a benefit because they enrolled a sicker group of patients because they mandated a higher dose of noradrenaline requirement before entry to the trial. So to look at that question, we did three things. We went back and looked at our predefined subgroups of sicker cohorts. We looked at geographic regions with a higher mortality. And three, we applied the APROX criteria to our trial to see if there would be any differences between the two. So when we looked at the sicker cohorts, we looked at the Apache score greater than 25. You can see the mortality rates are 38.8 versus 37.8%, no differences between the two groups. This is the adrenal cohort. When we looked at the severity of shock, we used a threshold of 15 mics per minute of noradrenaline, which translates to 0.22 mics per kilo per minute in a 70-kilogram individual. And again, the mortality differences were not significant, 33 versus 36%. We looked at geographic regions with a higher mortality, which was Denmark and Saudi and the UK, where the mortality rates were about 35% to close to 40%. And again, there were no differences between the two groups. If anything, the point estimate is in favor of placebo rather than hydrocortisone. And importantly, of course, the adrenal also affords generalizability because it was done on multiple sites, 69 sites across five different health care systems as a pragmatic trial. It was consistent in terms of day 90 and day 180 mortality, as you can see. We then applied the APROX criteria to the adrenal cohort, and 903 patients of the adrenal cohort met the APROX criteria. The mortality rates were very similar to the APROX cohort, 41 versus 44%, but the p-value confidence limits were wide p-value, not statistically significant. We then applied the sepsis-3 criteria because people say, well, you should have used sepsis-3. Well, the adrenal was conceived and designed before sepsis-3 was even thought of. And nevertheless, we went back and looked at our data set. 1,900 patients from the adrenal cohort met the sepsis-3 criteria. And again, the mortality rates were not different between the two. It's 32.4 versus 35.2, and a p-value not statistically significant. So we then reported these last year in anesthesiology, the results of the post-hoc analysis applying the APROX and the sepsis-3 criteria. Now, it's a useful opportunity to look at the four randomized controlled trials of low-dose corticosteroids. The GEDIN4-5 is, in fact, the original ANAND trial, the corticus, the adrenal, and the APROX. In the top row, you've got the sample size. And then you have the inclusion criteria, which are listed. And importantly, and one of the arguments that the French group made when corticus was published was that the corticus value, the results were not relevant because they had a high, large cohort of surgical patients as opposed to medical patients. So 66% was a 34% in the original French trial. But in fact, if you look at adrenal and APROX, it's predominantly medical sepsis in both trials. Now, you can see that this is the trial design, really. And the other key difference was that hydrocortisone was administered as an infusion in adrenal and as a bolus in APROX. And tapering was done in corticus. The other three trials did not use a tapering strategy for hydrocortisone. Now, these are the key outcome differences. Where it's highlighted in blue, they were not different where it says a Y in a gray background. And that is where there were differences. Now, interestingly, there were no mortality differences across GERIN45, corticus, and adrenal. In the intention to treat analysis, APROX observed a mortality difference in hospital, but not at day 28, but again, at day 90 and at six months, but not at 12 months. And again, what's interesting is that the survival benefit, which the French group reported in the original trial back in 2002 in the corticotropin non-responders, was not reproducible in the APROX trial. There was no survival advantage in the corticotropin non-responders administered hydrocortisone. And shock reversal was evident across all four trials. Earlier weaning was reported in adrenal and APROX. We reported a shorter ICU length of stay, and adverse effects were predominantly metabolic in adrenal and APROX. So these are the comparative sample sizes of low-dose steroids in septic shock. And as you can see, it's starting from 1998 to 2018. And the reason we arrive at a figure of 3,800 was the baseline mortality in septic shock has declined over the years. We aimed for an absolute risk reduction of 5%, aiming for the baseline mortality of 33%, and an alpha of 0.05 and a power of 0.9. The calculations came to 1,900 patients in each hour. Now, the advantage of doing this is it minimizes the risk of a type 1 error. And I've shown this here, especially for people who are planning clinical trials. So this is the DSMC. I've shown you the results of the interim analysis. We had three analyses. So the first one was done at 25% enrollment, where that's close to the unarmed cohort. Numbers there, 480 in each arm. And as you can see there, the mortality was 26% versus 31%. And the point estimate was in favor of hydrocortisone. It is a statistically significant p-value. But the DSMC said, no, you continue with the trial. You carry on. They applied the criteria. And they said, you carry on with the trial. And as the study progressed, at the 1,800 patient stage, and then the 2,500 patient stage, and then the final trial results, as you can see, the point estimate gradually moves towards the midline. And you lose the statistical significance. And this is very important, because if you stop the trial prematurely, because you've identified a treatment effect, there may well be a type 1 error. And that's something you've got to watch out for. So the summary message about low-dose steroids is that their role in improving mortality remains unclear. There are consistent improvement in important patient-centered secondary outcomes. And I believe the adrenal, and I'm sure people will debate that, but I believe the adrenal was a pragmatic trial, which was replicating real-world practice. We did not stipulate a minimum dose of preserves. Because if you looked at the clinical practice worldwide, people were using steroids at 2 mics per kilo of dopamine, 0.05 mics per kilo of noradrenaline. And in some situations, they were just using it for sepsis, not even septic shock. So it was not clear when, and therefore, what is the optimal threshold of catecholamines where institute steroids is not known. And that's what we did not stipulate. And so before the adrenal trial, I used to bulk when registrars used steroids for septic shock patients. Now I'm prepared to accept that if they are mechanically ventilated with septic shock and they require preserves for at least six hours, then I wouldn't object if they initiated corticosteroids. So the second question, of course, which comes out is, what's the role of flutricortisone? Now, could flutricortisone have made the difference? The basis of flutricortisone use is the notion that there is primary adrenal insufficiency in patients with septic shock. But the argument really is that hydrocortisone at 200 milligrams per day has sufficient mineralocorticoid activity. And therefore, endocrinologists would argue that you really do not need flutricortisone. So does it have any other independent vascular responsiveness effects? We don't know. Does it have any additional immunological effects? Might do. Again, it's not been clearly established. Now, three trials have investigated the role of flutricortisone. One is the original ANAND trial, which did not show any benefit on intention to treat, but only in a subgroup of corticotropin non-responders. The COITS trial, I'll show you, that was a two-by-two factorial design study, which did not show a benefit on intention to treat. And the APROX trial, which showed a benefit on intention to treat, but did not show a benefit in the corticotropin non-responders. And interestingly, in pharmacokinetic studies, where they've looked at pharmacoflutricortisone absorption in critically ill patients, they've not been able to demonstrate adequate serum levels. And that absorption has only been demonstrated in a third of the patients who are given nasogastric flutricortisone. So the only definitive way, if you want to study the role of flutricortisone, the only way is to do an RCT, which compares hydrocortisone versus hydrocortisone plus flutricortisone. Now, I told you this is the COITS trial. I told you this is the role of, which is a two-by-two factorial design study, which looked at conventional glucose control, hydrocortisone plus flutricortisone, the combination, and placebo. I just want to point out to you here that at least we know we've got data on shock reversal. Right at the bottom here, you've got hydro plus flutero there and hydrocortisone alone. The mortality difference is there between the two. The in-hospital was 42.9 in the combination and 45.8, but not statistically significant. And overall, mortality is 45.7 versus 48.5. And even in shock reversal, again, it did not make the median duration of vasopressor-free time. It was very similar between the two. Now, we used hydrocortisone as an infusion for logistic reasons. The APROX trial used placebo. Well, there really is much of a muchness, to be honest. And I think there is evidence that if you use it as an infusion, it attenuates the inflammatory response and the metabolic derangement is much less as an infusion as compared to a bolus. But there's really no clear evidence to support one mode or the other. What's the optimal duration of hydrocortisone treatment? Again, this is an unanswered question. In the ANAND trial, first trials, seven days. Corticose was seven days of hydrocortisone plus four days of tapering. That's 11 days. Adrenal was seven. APROX was seven. Now, I just want to show you, this is the hemodynamic graph in the adrenal trial. On the x-axis, you've got the time. On the y-axis, you've got the mean arterial pressure. And the red is hydrocortisone. The blue curve is placebo. And you can see a clear separation between the two curves, starting from day one, which is when the intervention commenced, up to day seven. And then the two curves converge once the intervention is stopped. So one could argue, had the intervention been continued beyond day seven, would there have been a separation of curves for a longer duration of time, which would have translated into improved organ function and outcomes? We don't know that. Is corticotropin testing required prior to administration of hydrocortisone? Again, I showed you that the only trial that showed that benefit was in the original ANAND trial, which has not been replicated subsequently. The corticotropin test has got limited utility in critical illness. And even a recent international task force meeting on adrenal insufficiency concluded that there is no reliable diagnostic test to diagnose adrenal insufficiency. Is it useful for preventing septic shock if you've got someone with a raging sepsis in your unit? And again, this was before sepsis 3 was developed. They tried this in patients with severe sepsis. And they identified that hydrocortisone really did not influence the progression to septic shock. Does the type of vasopressor influence the responsiveness to hydrocortisone? This question came up because in the original VAST trial, there was a suggestion that there was an interaction between vasopressin infusion and corticosteroid treatment. But this was subsequently not reproduced in the VANISH trial reported in 2016. Now, does the hydrocortisone require the presence of ascorbic acid and thiamine to be efficacious? And this was Paul Maddox's original report. And that was one of his key criticisms of the adrenal trial, that we did not use thiamine. Now, subsequently, we know the CITRUS-ALI trial did not show any significant benefit. The VITAMINS trial, more recently published in January, again did not show any benefit with VITAMIN C. And from adrenal, we've gone back and looked at the endogenous ascorbic acid concentrations. And we classify them into three tertiles. And from between the top and the bottom tertile, there were no differences in survival between the groups. So there's no evidence of synergy between ascorbic acid concentrations and response to hydrocortisone. If you have a patient with septic shock, and if they have a relapse after a course of hydrocortisone, should you restart hydrocortisone? Well, there's really no evidence to support or refute. So there's only one trial from Bruegel, which is really a small study. And I think you couldn't be criticized if you decided to give another course of hydrocortisone. So I think there are certain areas of future research which are important. I think it's got to be kept in mind that hydrocortisone has only been tested in high-income countries. In low- and middle-income countries, it's very important to look at this question because the pathogen profiles are different. The patients are often, they frequently have malnutrition, which can influence the immunological response and how that will interact with hydrocortisone, which is an immunosuppressive agent or something we don't know. Patients often present late in the piece. So I think this whole question needs to be looked at in that. Optimal duration is another question, and fluorocortisone is another key question. So we have summarized all these arguments in these publications. And I'll just quickly go through what we have done since we finished adrenal. We have looked at long-term outcomes, quality of life, which is Naomi's work. And essentially, overall, at six months, there were no differences between hydrocortisone and placebo when it came to these domains of mobility, self-care, usual activities, pain, discomfort, and anxiety and depression. They were very similar between the two groups. But what's interesting is the time to shock reversal had an influence in that if you reversed shock early, then fewer patients had problems with mobility, self-care, and usual activities, which was significant. So if it took longer than seven days to reverse shock, then there were problems. And so that's the key finding of the quality of life. We've done a large body of work on biology of vascular responsiveness. And so we have measured total and free cortisol and aldosterone concentrations given the interest in fluorocortisone and ascorbic acid. And I showed you the vitamin C data. And this is Jeremy Cohen's work, one of the authors of Adrenal. And we looked at the interaction with adjunctive hydrocortisone. And just again, to show you, this is the advantage of doing large trials. If you look at the sample size of previous studies of free cortisol, they barely got to about 70 or 80. And we've now got about 525 patients with plasma-free cortisol concentrations. And we've been able to show that, in fact, that free cortisol is better than total cortisol when it came to determining recurrence of shock, as a predictor of recurrence of shock. And when you look at, on the other hand, when you look at plasma-aldosterone, it had no predictive value in terms of shock resolution or shock reversal. But total and free cortisol did. And this is vitamin C. And again, so with the chair here, we have written an article which looked at the importance of the catecholamine-spreading strategies in patients with septic shock. And we have done some genomics. We have done a GWAS on 580 patients. We have done RNA sequencing on 580 patients. We have done some methylation work. And that will be reported soon. And we have an interesting, we have identified an association between a particular gene expression, a particular gene and vascular responsiveness. And we are quite excited by this and obviously waiting to report this. Now, has this changed practice two years down the track? Have people now, are they routinely prescribing steroids? Or are they still uncertain? Can I have a show of hands in the room here? How many prescribe corticosteroids for septic shock? And how many do not prescribe corticosteroids for septic shock? OK. So more hands went up for steroids. So I'll tell you what the international survey showed. So we did an international survey in 10 countries. And we had about, the survey went to about 500 people. And so the first question was, do you wait for a minimum duration of presotherapy before initiating? We had a number of questions. I'm just giving you the key ones. Before initiating hydrocortisone. And about two-thirds said yes, one-third said no. Do you wait for a minimum dose of presores before you initiate hydrocortisone therapy? And about 80% said yes and 20% said no. In patients to whom you prescribe hydrocortisone, do you also prescribe fluorocortisone? And only about 2% said yes always. And about 10% said sometimes. And the majority said no. So I guess there is still, I think the pendulum is swinging towards using steroid hydrocortisone. Fluorocortisone, I think there's still sufficient uncertainty there. So we are now working on an individual patient data meta-analysis with a French group. With the Pittsburgh group, we're looking at some precision medicine work. There's a health economic analysis in progress. And we're also looking at fluorocortisone dose-response relationships in a randomized controlled trial as a phase two study back in Brisbane. So this is a future research. So to conclude, the recommendations are, well, the practice recommendations. I think the role of hydrocortisone in improving mortality is unclear. There's clearly evidence of benefit in patient-centered secondary outcomes. No evidence of significant adverse effects. And so in mechanically ventilated patients with septic shock, my practice is to use hydrocortisone at 200 milligrams per day for seven days. You can use it for others infusion or bolus. There's no requirement for dose tapering, no requirement for a corticotropin test. Additional fluorocortisone is not necessary. Hydrocortisone can be used irrespective of which vasopressors being used. And it has no role in the prevention of septic shock. Thank you. Well, thank you, Bala. That was great. I'm going to open it up to the audience. But it looks like, from what I learned today, it's not yet time to end the steroids and septic shock debate, right? Look, yes. I guess, certainly when it comes to a mortality question, I think the debate still continues. And I think the other key question is, although we did this in five countries, that's all high-income countries, and it was generalizable to other parts of the world. So I think that question has to be addressed in LMICs as well. Yes, ma'am. Do you think a lot of the difference starts from the fact that the start of the steroids is variable? Therefore, the negative studies were because the institution of the treatment was very variable between practices, institutions, practitioners, and so forth. And maybe that made a difference. So the inclusion criteria for adrenal were that you had to, obviously, have septic shock, a pressure requirement for at least six hours, and be mechanically ventilated. In APROXX, the inclusion criteria were you have to have a catecholamine dosage of 0.25 mics per kilo per minute, and you could even reach that up to 24 hours later. So we do not actually know when they commenced hydrocortisone in their trial. So I think people need to be aware of that. And again, at the end of the day, when you do a clinical trial, these are some of the challenges we have. If you stipulate that steroids have got to be commenced within six hours of the patient coming to ICU, you probably won't be able to enroll many patients. And it also doesn't reflect real-world practice. This is what happened in the original ANAND trial. They started, they enrolled, their initial criteria was eight hours, and then they had to increase it to 24 hours because they just couldn't enroll any more patients. They modified their protocol. So these are the, well, and then again, when we looked at the time from onset of shock to randomization across the four different, in four six hourly epochs, we could not find any differences in outcome. Thank you. That was excellent. Thank you very much. I don't know if Dr. Sprung is here, but if Charlie were here, he'd know that I would say that his study, of course, was originally designed to enroll about 1,100 patients. 1,150, if I remember correctly, and he only enrolled about 500 and therefore never achieved the goal. And therefore, the biggest criticism I always made when that was presented at one of our meetings was that we really can't make a judgment about what that study said because he never enrolled as many patients as needed to demonstrate a difference if there was going to be one. I think that's important to remember. And you brought up errors in study design, whereas an ANAND study did reach its goal and take it for what it's worth. The results are what they are. Why wait? And I think this was the question I was just asking when I asked it again. Why wait at all? None of us wait to start pressors. And so my rule has been for 30 years, if you're starting pressors, and therefore have secondary shock, if you've resuscitated adequately with fluids first, and you're about to reach for pressors, well, then you start the steroids as well. Now, not all of my colleagues do that, but I do. And to your point, which I think is frankly crucial, pressor sparing ability or reduction in time of shock might be a very important goal in terms of ultimate outcomes that we've not explored as fully as we should. So I would say that should be the goal of a good big study such as yours. That was a tremendous and large study. That should be the goal. If you're starting pressors, that's one more thing. Give that to, start that now, give a dose. It doesn't have to be an infusion, as you point out. People do boluses, I do boluses. Why not? So good question. So the reason we specifically did not do that was because we were concerned that when patients were, so one of the requirements in our trial, entry requirements, was you had to be mechanically ventilated. So if patients received a bolus of propofol for intubation, and soon after intubation, they would be hypotensive, and then therefore they would initiate, and then the registrar's will commence pressors at that point of time. And so we did not want a sedation-induced hypotension as being diagnosed as septic shock. And so we just wanted that effect to be clear, and that's why we gave the six-hour period. So that, in reality, could you use it straight away? I mean, if you are convinced, or if they'd been on pressors for a while, if they'd been on pressors before they got intubated, then as soon as they got intubated, then we would give them corticosteroids straight away. So that's the reason for waiting the six-hour period. And to go back to, so, I mean, the corticus trial was, I sort of alluded to that briefly. They had to stop at 500. They planned to enroll 800 at the time, and two things happened. One was they ran out of the study drug, and B, the surviving sepsis campaign guidelines came out at the time, advocating a low-grade recommendation for the use of hydrocortisone. And therefore, clinicians felt that they would be legally liable if they withheld corticosteroids to patients with septic shock. And so they didn't have the equipoise, and so they had to stop the study. Yes, sir, over there. Hi, my name is Tutal. I'm a pharmacist over in California. Part of being a pharmacist, we always have to worry about drug costs, and in the US, we have a huge problem with increasing drug costs. For instance, vasopressin for our institution costs about 300 US per day, which is roughly about 450 Australian dollars, I believe. So I guess with what your practice has been, have you noticed, since it seems like you've been starting hydrocortisone sort of like within six hours of norepinephrine, have you noticed a decreased use of vasopressin or phenylephrine or epinephrine use? So we are doing a cost-economic analysis, and I'm hoping it'll show up, although the cost-economic methods, the analytical methods very often do not, they're not very granular, and they don't actually go down to that level of detail. For us, it's really based on an ICU bed day, and so everything gets factored in the cost of an ICU bed day, but you've raised a very, very good point there. And I think this is particularly relevant. Even if you do not find a mortality benefit, this is particularly relevant in low-middle-income countries because there, it's a user-paid service. Patients pay for every intervention. So if you can cut down the cost of vasopressin, if you can reduce the cost of intensive care, if you can reduce the duration of ICU, obviously, it'll have reduced the patient cost, plus it has implications for bed availability and triage. Will be the last two questions. Katrina Yenif from Henry Ford Hospital. How do you explain the differences in mortality between the control groups, the placebo groups? Most of ANAND studies have a 40% mortality in the placebo group. This is very similar to what we see from CMS data here in the United States. Your study, in conjunction with the ARISE study, had a really relatively low mortality, not necessarily 30-day mortality. It's a 90-day mortality, which is really low. How do you explain these differences? I cannot tell you why the mortality rate is different because this is our 33% mortality, which is what we used to calculate a sample size, was based on our survey in 2003, our sepsis survey, an epidemiological study, the way we observed a mortality of 33%. And subsequent interventional trials from ANZ have always had a low mortality. I believe this is a heart-worm effect we are seeing there? Do you believe it's better care provided overall to the general population? Look, I would like to be able to say so, but I'm sure the care is very comparable across regions. I think, but that's why we had to have a large sample to get the effect size of 5% from 33 to 28%. So our study was adequately powered, even for the low mortality. I can't explain any of the reason why we have a low mortality. Thank you. Rob. Rob McLaren from the University of Colorado. I apologize for my voice. I'm also a pharmacist. I'm always interested in the pharmacokinetics of drugs. And so with hydrocortisone, the half-life is six hours or so, plus or minus an hour. If I'm giving it as an infusion, it's not gonna get to steady state until it's 24, 30 hours later. So I'm just wondering why you didn't give a bolus up front whether you think maybe the infusion might've actually inhibited any benefit because now we're taking really 24 hours to see a therapeutic concentration. Well, your point is right. Yes, it'll take a much longer time to get to steady state. And we did not measure hydrocortisone concentrations, but we actually saw hemodynamic benefit. So if you see the hemodynamic traces, which I showed you, so as soon as the intervention was started, there was a clear separation between the two groups. So what is the optimal concentration required? We do not know. And the other thing, of course, is if you give it as a bolus, you get a large peak to start with, and then the concentration starts to go down pretty quickly. Ours is well-maintained. Once it achieves steady state, it's maintained throughout the seven days. So there are arguments for both, but the only way of determining that is to have plasma vial concentrations. That's great. Bala, actually, my mind is still slightly in the fog because I saw one slide of yours, your personal recommendation as we're wrapping up. It's be on mechanical ventilation and six hours of pressures, correct? So where does this mechanical ventilation piece fit into all of this? I will still certainly identify a cohort of septic shock patients where you could still be on some norepinephrine and not be ventilated. Is being on mechanical ventilation an absolute necessity in your books, or is this just a? No, so the reason we chose, I'm just essentially following the adrenal criteria. The reason we included mechanical ventilation as a requirement was because we have a number of patients with, say, urosepsis who come to ICU because they require pressures, which they can't be given in any other part of the hospital. Their mortality rates are very low, so we really want to include a cohort of patients who are likely to have a high risk of death. And therefore, we added mechanical ventilation in because that increases the risk of death. Once you have circulatory shock and respiratory failure, then the risk of death is higher and we want to choose the patients who would have a high risk of death. And hence, we built in mechanical ventilation. And now, based on those criteria, that's why I'm suggesting this. But yes, I mean, if people ended up with very high dose of pressures and even if they're not ventilated, we could still use it. So at least from the guidelines perspective, and I know this audience is still thinking, well, you know, if I pull this room, there'll probably still be a lot of people who wait for their second presser to go on before they do their hydrocortisone, right? And I think it's the area of the big question mark is how much norepinephrine and what duration? Well, you said six hours, but how much norepinephrine? And there's all of us who switched to vaso at different levels of norepinephrine. So there's still a certain element of uncertainty and practice uncertainty, and probably the surviving sepsis guidelines need to start commenting on all of that as well, because it looks like it's an important element of our care. Sure, absolutely. So if you, I don't think you need to think through a threshold of noradrenaline before you start steroids because we did not mandate a threshold dosage, and worldwide, the practices vary, and that's why, to make it to be pragmatic, that's why we, six hours of threshold requirement, you start steroids. If you go by the APROX trial, then you'll have to have 0.25 mics per kilo of noradrenaline before you can score a dose of steroids, and that's too high. You can't do that. So you've got to draw a line some way, and I think that it's very reasonable, and you will have sufficient grounds to justify it if you decide to use steroids if a person has a duration of septic shock which exceeds whatever, four to six hours. I think that's very reasonable to do that. Well, thank you again. Thank you to the audience. I hope a lot of questions are answered. Otherwise, he's available afterwards as well. Thank you. Thank you all.

steroids

septic shock

sepsis

dosing strategies

adrenal trial

hydrocortisone

mortality

mechanical ventilation

recommendations

Pharmacology

Shock

Inflammation

Presentation

Select

Professional

Associate

2020

Advanced

Sepsis Resources