The keys to sepsis care are simple: (1) high index of suspicion; (2) early, broad-spectrum antibiotics; and (3) aggressive volume resuscitation in those who are volume depleted/responsive, vasopressors for those who need them. That’s it. The SEP-1 bundle makes proscribes specific things that clearly don’t improve outcomes, and now, it will penalize us if we don’t comply. It’s ridiculous.
And, just so you know, my hospital has a greater than 89% bundle compliance and fantastic sepsis outcomes. I’m on the sepsis committee. So I’m not complaining because we suck at sepsis.
I feel like this is spot on. Although EGDT was disproven it’s like no one ever let it go. And it lives on in SEP-1. The elements of care we can control are very limited. But that’s hard to admit and accept.
Exactly … Once protocols get written into what is effectively law, all sorts of strange knock-on effects happen. I agree there needs to be some type of measurement, but sepsis is inherently resistant to standardization of care.
The fact that the agencies voicing support are from critical care and the voices against are from emergency medicine and ID, is very relevant. I can comment on both environments (ER and ICU) since I practice in both. How these "guidelines" (in quotes for a reason) play out in practicality is very different from what you might realize if you do not actually work in a busy ER. To begin with, these guidelines are unbelievably non-specific. Almost anyone with any kind of stress response whatsoever meets these criteria, and if the patient has a fever then most will reasonably suspect an infectious source, meaning the patient is "septic". And, most patients with fever have some amount of tachycardia.
The idea that ALL of them require EXACTLY 30 cc/kg immediately is patently absurd. Fluid requirements are very different for different patients. Giving that amount of fluid as a bolus to someone in florid pulmonary edema is very dangerous and very harmful. On the flip side, stopping after 30 cc/kg to the drop (literally) and starting vasopressors when they still need more fluids is almost equally as bonkers. Who seriously thinks this is the right thing to do?
The idea that everyone spiking a fever in the ER gets broad spectrum abx without even knowing the source is a bad idea for many reasons, not the least of which is the development of abx resistance.
Think we are allowed to exercise our judgement in the ER? False. When hospitals lose millions for not meeting these "bundles" they are no longer guidelines. They are laws. Where else in medicine are clinicians mandated to go against their judgement and do what they feel is harmful to the patient in front of them.
At many hospitals, like mine, we have to call a "code sepsis" overhead, alerting/waking the whole hospital for every patient presenting with these criteria. And if you have not given 30 cc/kg bolus, or given broad spectrum abx, you are "dinged" by the nurse with a clipboard reviewing your chart in a few days.
This is madness. And it seems to be getting worse and worse. Again, if you do not work in the ER you are not fully appreciating the effects this is having. If you only look at those who end up down the road having a diagnosis of severe sepsis, that is ignoring the countless patients who were harmed directly or indirectly (because the staff was tied up in a pointless "code sepsis" for a different patient while you yourself were crashing) by these nonsensical protocols.
I signed up for PCCM specifically to address this particular article and the issues it purports to discuss.
I. First, the article begins with Q-Anon-like inuendo and conspiracy theories that are not germane to SEP-1. These theories were exposed as factually incorrect more than a decade ago. The financial arrangements between Henry Ford, Manny Rivers, and Edwards are exactly the same as what might be expected for a clinical trial. Name the drug you use in your practice that did not pay hospitals and local PIs for running its clinical trials. Tell us precisely why that should not be true for Edwards’ catheters. In fact, the ~ $350,000 dollars of his “own” money, i.e. money held at Henry Ford for his use, that Dr. Rivers spent on the trial were earned by participating in a number of pharmaceutical company therapeutic trials for sepsis. No one complains about that. (Ann Emerg Med 52:651-653, 2008.)
II. The person who reported irregularities in the Rivers trial has not actually “come forward” and remains un-named. Whoever this person was seems to not be familiar with clinical trial procedures. The claim is that 25 control patients were dropped from the trial, which happens to be the same number of patients that Dr. Rivers reported as screened for the trial but not randomized because they failed to meet study criteria or did not consent to participation. This is standard for any clinical trial.
III. The trio of EGDT papers that supposedly debunked the process were massively flawed and were designed in such a way that they could not even test the actual hypothesis. Why do I say this? Recall that the goal of EGDT is to quickly achieve adequate tissue oxygenation, as evidenced by an ScvO2 > 70%. The median ScvO2 in all three of ARISE, ProCESS and ProMISE was > 70%. In other words, half of the randomized patients already met the goal at the time of randomization. None of the studies comments on what steps were taken in those patients. However, it is clear that at least half of patients in the studies did not need EGDT, hence the studies were all under-powered to find a difference in outcomes. None of them found a difference in outcomes. I’m not going to say that I still used EGDT, but I do, in fact, check ScvO2 in any septic shock patient who gets a central line. Of course, that is far fewer than it used to be, since we now know that vasopressors can be safely delivered peripherally in many cases.
IV. SEP-1 is not EGDT and never was. One might consider it a step-child, at best, or possibly a grandchild of EGDT, if one steps back and squints. In the first place, EGDT only ever applied to septic shock, and not to sepsis or severe sepsis. I use those terms, because CMS continues to do so. Secondly, measuring lactate and following it once if it is increased on the first measure is perhaps somewhat related to EGDT, but was never a component of EGDT. Rapid administration of antibiotics was never a specific component of EGDT. Drawing blood cultures before antibiotics was never a component of EGDT. One component of SEP-1 that descends from the EGDT trial is the 30 mL/kg for septic shock patients, which was partially derived from what patients received in the Rivers trial.
V. The most controversial aspect of SEP-1 is the 30 mL/kg fluid requirement. However, 30 mL/kg is NOT a requirement. Many reasons for not administering it are acceptable to CMS, as long as the reason is given in the medical record. “The patient has pulmonary edema”, would work. “The patient has a history of heart failure”, would work. “The patient failed to respond to a passive leg raise with increased stroke volume or increased cardiac output or increased pulse pressure”, would work. Lots of legitimate reasons would work. People who complain about this aspect of SEP-1 are mostly complaining either because they don’t know it or because they don’t want to record their thoughts in the EMR. Any malpractice lawyer can tell you that isn’t wise. It’s your thoughts as a physician on the spot that are the most important thing to record in the medical record, even though we’re all aware that billing requirements these days dictate a lot of junk must be recorded.
VI. The recent “systematic review” of SEP-1’s effect on mortality amassed quite a number of studies that were poorly designed to answer the question. Most of those studies took an interrupted time series approach and asked “what was sepsis mortality before SEP-1 reporting became a rule vs. after”. This design is deeply flawed and can’t answer the question of whether SEP-1 works. All but one of the studies fails to address how compliant the hospitals were with the measures after their institution. SEP-1 measures cannot possibly make a difference if they are not being used. Or, from the other side, none of the studies addresses how many of the measures were actually already being followed before SEP-1 came to be. Either way, if practices have not changed, then the presence of a reporting requirement makes absolutely zero difference. I’m sure most reader here drive a car. Do they all adhere strictly to posted speed limits? All the time? Why not? Because most of the time there is no consequence. I think you get my point. An interrupted time series done now that there is financial consequence and comparing 2015 – 2024 data with later might have some hope of answering the question. So far, the best data on SEP-1’s effectiveness come from CMS, are well characterized and balanced for confounders, and demonstrate benefit. Combining a bunch of bad studies into a “systematic review” does not give them any more credence.
VII. It is difficult to say how many hospitals in the US had a sepsis quality committee previous to 2015, the inception of SEP-1. However, in 2022, seven years after the advent of SEP-1, according to the CDC only 73% of hospitals had a sepsis committee, so it’s likely that fewer than half did before SEP-1 went live. Will you really posit that hospitals should not pay attention to sepsis, which has been demonstrated over and over to be the most deadly and most expensive condition treated in US hospitals? Addition to value based purchasing (VBP) will only increase the number of people paying attention.
VIII. Back to ARISE, ProCESS and ProMISE. They all demonstrate that standard care in high quality centers has migrated to early fluid and antibiotic administration. “Usual care” means something more than what it once did. It may well be that usual care has progressed to the point that SEP-1 is now behind the times. CMS and CDC have committed to developing outcomes measures. If you think that VBP of the current SEP-1 is a problem, just wait until your hospital is financially dinged because your outcomes are in the bottom 10th percentile. And it’s difficult to argue that you shouldn’t be.
IX. I will accede here that hospitals CAN have quite good sepsis outcomes while adhering poorly to the all-or-none SEP-1 construct, and the outcomes based assessment may well make sense. There is logic to believing that missing, say, antibiotic timing by 15 minutes here and there, or failing to document why one did not give 30 mL/kg to an already fluid overloaded patient, or forgetting to get blood cultures before giving antibiotics on occasion won’t affect your outcomes tremendously. However, put on your hat as a septic patient, which I have been, or the family of a septic shock patient, which I have also been, and ask yourself whether your local hospital should not have at least a floor for expectation of what they should do. As chair of the board of Sepsis Alliance (www.sepsis.org), I can show you many stories on our website of patients whose hospitals did not have this floor and who did not fare so well. I was lucky that my dad’s local hospital complied with the SEP-1 bundle, even while they tried to convince me that his shock was cardiogenic, because his troponin was mildly increased. Never mind that he had an obvious infection source and his EKG showed no evidence of STEMI. In my n of 1 study, the presence of the bundle made a difference.
X. As the former president of what you call the “chest physicians”, I would ask that you just refer to the organization as CHEST, or as the American College of Chest Physicians. The American Thoracic Society has not commented on SEP-1, but endorses the 2021 Surviving Sepsis Guidelines, which are identical in early treatment of sepsis to SEP-1. I will point out that the national societies who militate against SEP-1 are either those whose members are held responsible for carrying out most of the SEP-1 measures in emergency departments or those whose members are not on the front lines of early sepsis care.
XI. I have to agree that RCTs can/should be of use. At least one is underway (NCT05491941) Assessment for Implementation Methods in Sepsis (AIMS). However, this trial really is comparing antibiotic delivery in one hour vs three hours. The SEP-1 bundle elements are included in both arms of the trial. I don’t think that is the trial you are asking for, but I have to wonder what you would want to compare SEP-1 with, given that its measures are actually pretty standard, anyway,
Your approach in this treatise feels derogatory, but it is based in literature that is flawed and that it doesn’t seem that you have read with a critical eye. Critical thinking is important on both sides of this issue.
The keys to sepsis care are simple: (1) high index of suspicion; (2) early, broad-spectrum antibiotics; and (3) aggressive volume resuscitation in those who are volume depleted/responsive, vasopressors for those who need them. That’s it. The SEP-1 bundle makes proscribes specific things that clearly don’t improve outcomes, and now, it will penalize us if we don’t comply. It’s ridiculous.
And, just so you know, my hospital has a greater than 89% bundle compliance and fantastic sepsis outcomes. I’m on the sepsis committee. So I’m not complaining because we suck at sepsis.
I feel like this is spot on. Although EGDT was disproven it’s like no one ever let it go. And it lives on in SEP-1. The elements of care we can control are very limited. But that’s hard to admit and accept.
This has inspired another podcast episode. Will post it once it's done.
Simply put..enforced reliance on protocol based medical decision trees is lousy medicine and has greatly diminished physician led healthcare.
There are so many unforeseen ramifications. None of which are good for patient care.
Exactly … Once protocols get written into what is effectively law, all sorts of strange knock-on effects happen. I agree there needs to be some type of measurement, but sepsis is inherently resistant to standardization of care.
The fact that the agencies voicing support are from critical care and the voices against are from emergency medicine and ID, is very relevant. I can comment on both environments (ER and ICU) since I practice in both. How these "guidelines" (in quotes for a reason) play out in practicality is very different from what you might realize if you do not actually work in a busy ER. To begin with, these guidelines are unbelievably non-specific. Almost anyone with any kind of stress response whatsoever meets these criteria, and if the patient has a fever then most will reasonably suspect an infectious source, meaning the patient is "septic". And, most patients with fever have some amount of tachycardia.
The idea that ALL of them require EXACTLY 30 cc/kg immediately is patently absurd. Fluid requirements are very different for different patients. Giving that amount of fluid as a bolus to someone in florid pulmonary edema is very dangerous and very harmful. On the flip side, stopping after 30 cc/kg to the drop (literally) and starting vasopressors when they still need more fluids is almost equally as bonkers. Who seriously thinks this is the right thing to do?
The idea that everyone spiking a fever in the ER gets broad spectrum abx without even knowing the source is a bad idea for many reasons, not the least of which is the development of abx resistance.
Think we are allowed to exercise our judgement in the ER? False. When hospitals lose millions for not meeting these "bundles" they are no longer guidelines. They are laws. Where else in medicine are clinicians mandated to go against their judgement and do what they feel is harmful to the patient in front of them.
At many hospitals, like mine, we have to call a "code sepsis" overhead, alerting/waking the whole hospital for every patient presenting with these criteria. And if you have not given 30 cc/kg bolus, or given broad spectrum abx, you are "dinged" by the nurse with a clipboard reviewing your chart in a few days.
This is madness. And it seems to be getting worse and worse. Again, if you do not work in the ER you are not fully appreciating the effects this is having. If you only look at those who end up down the road having a diagnosis of severe sepsis, that is ignoring the countless patients who were harmed directly or indirectly (because the staff was tied up in a pointless "code sepsis" for a different patient while you yourself were crashing) by these nonsensical protocols.
PLEASE CAN WE STOP!
AMEN
I signed up for PCCM specifically to address this particular article and the issues it purports to discuss.
I. First, the article begins with Q-Anon-like inuendo and conspiracy theories that are not germane to SEP-1. These theories were exposed as factually incorrect more than a decade ago. The financial arrangements between Henry Ford, Manny Rivers, and Edwards are exactly the same as what might be expected for a clinical trial. Name the drug you use in your practice that did not pay hospitals and local PIs for running its clinical trials. Tell us precisely why that should not be true for Edwards’ catheters. In fact, the ~ $350,000 dollars of his “own” money, i.e. money held at Henry Ford for his use, that Dr. Rivers spent on the trial were earned by participating in a number of pharmaceutical company therapeutic trials for sepsis. No one complains about that. (Ann Emerg Med 52:651-653, 2008.)
II. The person who reported irregularities in the Rivers trial has not actually “come forward” and remains un-named. Whoever this person was seems to not be familiar with clinical trial procedures. The claim is that 25 control patients were dropped from the trial, which happens to be the same number of patients that Dr. Rivers reported as screened for the trial but not randomized because they failed to meet study criteria or did not consent to participation. This is standard for any clinical trial.
III. The trio of EGDT papers that supposedly debunked the process were massively flawed and were designed in such a way that they could not even test the actual hypothesis. Why do I say this? Recall that the goal of EGDT is to quickly achieve adequate tissue oxygenation, as evidenced by an ScvO2 > 70%. The median ScvO2 in all three of ARISE, ProCESS and ProMISE was > 70%. In other words, half of the randomized patients already met the goal at the time of randomization. None of the studies comments on what steps were taken in those patients. However, it is clear that at least half of patients in the studies did not need EGDT, hence the studies were all under-powered to find a difference in outcomes. None of them found a difference in outcomes. I’m not going to say that I still used EGDT, but I do, in fact, check ScvO2 in any septic shock patient who gets a central line. Of course, that is far fewer than it used to be, since we now know that vasopressors can be safely delivered peripherally in many cases.
IV. SEP-1 is not EGDT and never was. One might consider it a step-child, at best, or possibly a grandchild of EGDT, if one steps back and squints. In the first place, EGDT only ever applied to septic shock, and not to sepsis or severe sepsis. I use those terms, because CMS continues to do so. Secondly, measuring lactate and following it once if it is increased on the first measure is perhaps somewhat related to EGDT, but was never a component of EGDT. Rapid administration of antibiotics was never a specific component of EGDT. Drawing blood cultures before antibiotics was never a component of EGDT. One component of SEP-1 that descends from the EGDT trial is the 30 mL/kg for septic shock patients, which was partially derived from what patients received in the Rivers trial.
V. The most controversial aspect of SEP-1 is the 30 mL/kg fluid requirement. However, 30 mL/kg is NOT a requirement. Many reasons for not administering it are acceptable to CMS, as long as the reason is given in the medical record. “The patient has pulmonary edema”, would work. “The patient has a history of heart failure”, would work. “The patient failed to respond to a passive leg raise with increased stroke volume or increased cardiac output or increased pulse pressure”, would work. Lots of legitimate reasons would work. People who complain about this aspect of SEP-1 are mostly complaining either because they don’t know it or because they don’t want to record their thoughts in the EMR. Any malpractice lawyer can tell you that isn’t wise. It’s your thoughts as a physician on the spot that are the most important thing to record in the medical record, even though we’re all aware that billing requirements these days dictate a lot of junk must be recorded.
VI. The recent “systematic review” of SEP-1’s effect on mortality amassed quite a number of studies that were poorly designed to answer the question. Most of those studies took an interrupted time series approach and asked “what was sepsis mortality before SEP-1 reporting became a rule vs. after”. This design is deeply flawed and can’t answer the question of whether SEP-1 works. All but one of the studies fails to address how compliant the hospitals were with the measures after their institution. SEP-1 measures cannot possibly make a difference if they are not being used. Or, from the other side, none of the studies addresses how many of the measures were actually already being followed before SEP-1 came to be. Either way, if practices have not changed, then the presence of a reporting requirement makes absolutely zero difference. I’m sure most reader here drive a car. Do they all adhere strictly to posted speed limits? All the time? Why not? Because most of the time there is no consequence. I think you get my point. An interrupted time series done now that there is financial consequence and comparing 2015 – 2024 data with later might have some hope of answering the question. So far, the best data on SEP-1’s effectiveness come from CMS, are well characterized and balanced for confounders, and demonstrate benefit. Combining a bunch of bad studies into a “systematic review” does not give them any more credence.
VII. It is difficult to say how many hospitals in the US had a sepsis quality committee previous to 2015, the inception of SEP-1. However, in 2022, seven years after the advent of SEP-1, according to the CDC only 73% of hospitals had a sepsis committee, so it’s likely that fewer than half did before SEP-1 went live. Will you really posit that hospitals should not pay attention to sepsis, which has been demonstrated over and over to be the most deadly and most expensive condition treated in US hospitals? Addition to value based purchasing (VBP) will only increase the number of people paying attention.
VIII. Back to ARISE, ProCESS and ProMISE. They all demonstrate that standard care in high quality centers has migrated to early fluid and antibiotic administration. “Usual care” means something more than what it once did. It may well be that usual care has progressed to the point that SEP-1 is now behind the times. CMS and CDC have committed to developing outcomes measures. If you think that VBP of the current SEP-1 is a problem, just wait until your hospital is financially dinged because your outcomes are in the bottom 10th percentile. And it’s difficult to argue that you shouldn’t be.
IX. I will accede here that hospitals CAN have quite good sepsis outcomes while adhering poorly to the all-or-none SEP-1 construct, and the outcomes based assessment may well make sense. There is logic to believing that missing, say, antibiotic timing by 15 minutes here and there, or failing to document why one did not give 30 mL/kg to an already fluid overloaded patient, or forgetting to get blood cultures before giving antibiotics on occasion won’t affect your outcomes tremendously. However, put on your hat as a septic patient, which I have been, or the family of a septic shock patient, which I have also been, and ask yourself whether your local hospital should not have at least a floor for expectation of what they should do. As chair of the board of Sepsis Alliance (www.sepsis.org), I can show you many stories on our website of patients whose hospitals did not have this floor and who did not fare so well. I was lucky that my dad’s local hospital complied with the SEP-1 bundle, even while they tried to convince me that his shock was cardiogenic, because his troponin was mildly increased. Never mind that he had an obvious infection source and his EKG showed no evidence of STEMI. In my n of 1 study, the presence of the bundle made a difference.
X. As the former president of what you call the “chest physicians”, I would ask that you just refer to the organization as CHEST, or as the American College of Chest Physicians. The American Thoracic Society has not commented on SEP-1, but endorses the 2021 Surviving Sepsis Guidelines, which are identical in early treatment of sepsis to SEP-1. I will point out that the national societies who militate against SEP-1 are either those whose members are held responsible for carrying out most of the SEP-1 measures in emergency departments or those whose members are not on the front lines of early sepsis care.
XI. I have to agree that RCTs can/should be of use. At least one is underway (NCT05491941) Assessment for Implementation Methods in Sepsis (AIMS). However, this trial really is comparing antibiotic delivery in one hour vs three hours. The SEP-1 bundle elements are included in both arms of the trial. I don’t think that is the trial you are asking for, but I have to wonder what you would want to compare SEP-1 with, given that its measures are actually pretty standard, anyway,
Your approach in this treatise feels derogatory, but it is based in literature that is flawed and that it doesn’t seem that you have read with a critical eye. Critical thinking is important on both sides of this issue.
I appreciate the thoughtful response, Dr. Simpson.