Beta-blockers and septic shock

This forum made possible through the generous support of SDN members, donors, and sponsors. Thank you.

lymphocyte

Full Member
7+ Year Member
Joined
Feb 27, 2015
Messages
2,265
Reaction score
4,219
MS4. Sorry in advance if inappropriate or dumb. I'm just curious if/how/how often you're using beta-blockers in septic shock.

What prompted this: there was an interesting back-and-forth last month in ICM. It had to do with another Morelli paper on physiological parameters in a cohort of 45 patients in septic shock and tachy at 24hr. They got esmolol in a HR-directed fashion. Arterial elastance improved, SV increased, CO unchanged. Pressor requirements dropped. Seems in line with past literature.

Lots of subgroups to consider (cardiomyopathy, ventricular dyskinesia, etc.), but in general is your practice changing? Maybe? Apparently, there are two big trials underway in France and Beth Israel Deaconess.

Last thread I could I find was in 2012.

http://forums.studentdoctor.net/threads/beta-blockers-and-norepinephrine-in-septic-shock.961551/

Members don't see this ad.
 
Last edited:
  • Like
Reactions: 1 user
This is paper you are probably referring too.

http://jama.jamanetwork.com/article.aspx?articleid=1752246

This should not be extrapolated to all comers. I treat pediatric patients and not adults and so I don't have expertise in this specific population (no one is going to give B-blockers to pediatric patients even though they have higher resting heart rates), but a closer observation of the data reveals several reasons why one should not interpret the use of beta-blockers in sepsis as a good thing. First and foremost, one single center study of 77 patients should never be extrapolated to a population as whole. This has created so many problems in the critical care world (and in clinical research) in general, where people take the results of a single center study and adopt the treatment/protocol/what have you to all patients (eg. EDGT therapy in 2001), but when it gets repeated in a multi-center fashion (EGDT trials of 2014 and 2015), the original findings are refuted or at least not shown to be as beneficial as one initially thought. This has lead people to move away from single center studies as being definitive for change in practice (see http://www.ncbi.nlm.nih.gov/pubmed/19789447).

Second, the mortality in the placebo group was 80%! Now, even in the most recent NEJM trials of EDGT versus controls, the mortality was around 25-30% for both groups. The original EDGT trial of 2001 had a mortality of 56%. 80% mortality in sepsis is exceptionally unusual nowadays and is typically quoted in the 30 to 50% range (probably on the lower end). Even the Esmolol group in this trial barely achieved what most would consider a normal rate of survival. So, its really hard to say a therapy is beneficial to a global septic population when the controls or intervention don't reach what is considered to be an acceptable range of mortality. I instead interpret the data as this center is really bad at taking care of septic patients and if you have sepsis in Rome, you need to drive to the nearest vineyard and sample as much wine as possible in the time you have left because either way, you are certainly going to die.

Likely, more studies would be needed to confirm this finding (like the ones you mention) and I suspect, that applied broadly, this study will be just like every other study in sepsis with a single intervention, that being one that showed initial improvement in one center but be showed no benefit in a multi-center approach.
 
Last edited:
  • Like
Reactions: 2 users
Members don't see this ad :)
I believe the answer is:

Sepsis (and by extension) sepsis treatment is not a one size fit all category. There is likely to be many different phenotypes of sepsis probably based upon some combination of specific organism, genetics, underlying physiology, and probably many other things. Different treatments, such as steroids or b-blockers as an example, probably work for certain patients but not for others.


As to actually identifying who possesses those phenotypes however... Still workin on it.
 
  • Like
Reactions: 1 user
This is paper you are probably referring too.

http://jama.jamanetwork.com/article.aspx?articleid=1752246

This should not be extrapolated to all comers. I treat pediatric patients and not adults and so I don't have expertise in this specific population (no one is going to give B-blockers to pediatric patients even though they have higher resting heart rates), but a closer observation of the data reveals several reasons why one should not interpret the use of beta-blockers in sepsis as a good thing. First and foremost, one single center study of 77 patients should never be extrapolated to a population as whole. This has created so many problems in the critical care world (and in clinical research) in general, where people take the results of a single center study and adopt the treatment/protocol/what have you to all patients (eg. EDGT therapy in 2001), but when it gets repeated in a multi-center fashion (EGDT trials of 2014 and 2015), the original findings are refuted or at least not shown to be as beneficial as one initially thought. This has lead people to move away from single center studies as being definitive for change in practice (see http://www.ncbi.nlm.nih.gov/pubmed/19789447).

Second, the mortality in the placebo group was 80%! Now, even in the most recent NEJM trials of EDGT versus controls, the mortality was around 25-30% for both groups. The original EDGT trial of 2001 had a mortality of 56%. 80% mortality in sepsis is exceptionally unusual nowadays and is typically quoted in the 30 to 50% range (probably on the lower end). Even the Esmolol group in this trial barely achieved what most would consider a normal rate of survival. So, its really hard to say a therapy is beneficial to a global septic population when the controls or intervention don't reach what is considered to be an acceptable range of mortality. I instead interpret the data as this center is really bad at taking care of septic patients and if you have sepsis in Rome, you need to drive to the nearest vineyard and sample as much wine as possible in the time you have left because either way, you are certainly going to die.

Likely, more studies would be needed to confirm this finding (like the ones you mention) and I suspect, that applied broadly, this study will be just like every other study in sepsis with a single intervention, that being one that showed initial improvement in one center but be showed no benefit in a multi-center approach.

I tried to hyperlink (fail) the new Morelli paper from 2016:

http://link.springer.com/article/10.1007/s00134-016-4351-2

I keep saying "Morelli" because it's not just single center, it's the same lead author and non-blinded, so I'm naturally skeptical. Very strongly agreed about the Rivers' trial and all the lessons to be learned from it.

One challenge I've noticed with intensive medicine is adequately powering trials, since each patient is so nuanced and tricky. That makes physiological arguments a little more attractive, like with this new paper. I was wondering if these arguments were convincing anybody. But "not quite yet" is the sense I'm getting and fairly so.

ESMOSEPSIS from France should be out next year, so I guess we'll have more data then. But it's still open-label, industry-sponsored, with the primary outcome being cardiac index, and the secondary outcomes being pressor requirements and measures of microcirculatory flow--not exactly the hardest of end-points. Cocchi is running the study at BID with a completion date of 2019. Either way, it's going to be a long wait till 1) we get useful mortality data, and 2) any kind of respectable power across the hyperkinetic septic spectrum.
 
Last edited:

Just reading over the last thread, I guess all there is now is more physiological ammunition to argue with. But those in favour of esmolol have a lot more ammo: improved arterial elastance and unchanged CO, despite a drop in pressor requirements. Plus, you can always switch it off...

That 2013 study really bothered me:

1. SAPS II didn't predict mortality in the controls.
2. Only powered to detect change in HR.
3. 154 patients with umpteen variables in the Cox model.
4. Open label with declared conflicts-of-interests (yes, sorry for the tinfoil hat, but still).
5. Most importantly, everybody focused on mortality--that was the show-stopper and a source for a lot of skepticism--but I thought the real strength was clinically meaningful surrogates like GFR. Mortality is not a great indicator in small-N trials for sick, heterogenous patients.
 
Last edited:
One challenge I've noticed with intensive medicine is adequately powering trials, since each patient is so nuanced and tricky. That makes physiological arguments a little more attractive, like with this new paper. I was wondering if these arguments were convincing anybody. But "not quite yet" is the sense I'm getting and fairly so.

Trying to correct physiology is a fine goal, but just engineerdout mentioned, there are nearly an infinite number of roads that led to the physiology being deranged (organism, timing/onset, co-morbidites, genomics, proteomics) and maybe beta-blockers address some subset, but considering there are 1.4 million admissions for sepsis in the US per year, that subset 1) is impossible to determine prospectively at this point 2) is likely so small that applying a therapy broadly either has no benefit, or worse, causes harm. Some of the genomics studies in sepsis retrospectively shown that a certain genomic profile leads to a good outcome when steroids are used, and another genomic profile leads to death when steroids are used (and this is a drug that has common acceptance in sepsis... god knows why). I suspect the same is true for whatever single therapy one wants to employ. Unfortunately, most genomic/proteomic studies which employ GWAS or RNAseq take about 1 to 2 months to get information back. By that point the patient either survived or didn't.

As far as that future study you pointed to, it just continues to show how naive clinical studies are at addressing complex pathophysiology. One of the goals of the study is to show how beta-blockers cause "Changes in the Cytokine pattern(pro and anti inflammatory cytokines assay)". You can get into the esoterics of adrenergic stimulation and it effects on immune system, but they are including this not because they have a testable hypothesis, they are doing because they can and I guarantee you if they get any p <0.05, they will spin the argument in the direction that is most likely to lead to the highest impact factor publication. All the while, the results are pretty meaningless because despite what people want to believe, correlation is not the same as cause and effect.
 
Last edited:
  • Like
Reactions: 1 user
Just reading over the last thread, I guess all there is now is more physiological ammunition to argue with. But those in favour of esmolol have a lot more ammo: improved arterial elastance and unchanged CO, despite a drop in pressor requirements. Plus, you can always switch it off...

That 2013 study really bothered me:

1. SAPS II didn't predict mortality in the controls.
2. Only powered to detect change in HR.
3. 154 patients with umpteen variables in the Cox model.
4. Open label with declared conflicts-of-interests (yes, sorry for the tinfoil hat, but still).
5. Most importantly, everybody focused on mortality--that was the show-stopper and a source for a lot of skepticism--but I thought the real strength was clinically meaningful surrogates like GFR. Mortality is not a great indicator in small-N trials for sick, heterogenous patients.

This really isn't a patient centered outcome. Unless these patients end up with a less long-term CKD, lower mortality, or something else that matter who cares if you affect the GFR slightly for a small period of time.

This is an antidote, but the "cleanest kill" in medicine I saw involved a beta-blocker in sepsis. A nurse was concerned about a tachy septic patient on multiple pressors and asked for medication to "treat the heart rate." Intern gives some lopressor. 5 minutes later patient is pulseless and code blue is called. 15 minutes later time of death is called. Patients with crappy hearts need a high heart rate during sepsis to maintain blood pressure. Taking that away can be fatal.
 
  • Like
Reactions: 2 users
This really isn't a patient centered outcome. Unless these patients end up with a less long-term CKD, lower mortality, or something else that matter who cares if you affect the GFR slightly for a small period of time.

This is an antidote, but the "cleanest kill" in medicine I saw involved a beta-blocker in sepsis. A nurse was concerned about a tachy septic patient on multiple pressors and asked for medication to "treat the heart rate." Intern gives some lopressor. 5 minutes later patient is pulseless and code blue is called. 15 minutes later time of death is called. Patients with crappy hearts need a high heart rate during sepsis to maintain blood pressure. Taking that away can be fatal.

1. Beta 1 vs beta 2 vs both. And pharmacokinetics of metoprolol vs esmolol. And HR-directed therapy (or some other reasonable endpoint) after haemodynamic stabilisation. And MAP is more than just HR: SV increased in the Morelli paper (though CO was unchanged... though arterial elastance improved). And isn't cardiac perfusion mostly diastolic anyways? And isn't the pressor still helping with MAP via alpha-1? (Stronger argument that beta-2 blockade will harm MAP.)

2. There are other lines of evidence that catecholamine excess might be harmful too, like in certain sub-groups of CHF or stress cardiomyopathy--but it's always dangerous to draw pathophysiological analogies, and I appreciate that.

3. Patient-centred outcomes take a long time to adequately power, especially with sick patients, so isn't it fair to become responsive to meaningful physiological signals of benefit or harm (like with GFR)? It's like with NICE-SUGAR and the Leuvan trials. (Which I understand also burned early adopters and rightfully suggests caution here--but that was a long seven years!).
 
Last edited:
1. Beta 1 vs beta 2 vs both. And pharmacokinetics of metoprolol vs esmolol. And HR-directed therapy (or some other reasonable endpoint) after haemodynamic stabilisation. And MAP is more than just HR: SV increased in the Morelli paper (though CO was unchanged... though arterial elastance improved). And isn't cardiac perfusion mostly diastolic anyways? And isn't the pressor still helping with MAP via alpha-1? (Stronger argument that beta-2 blockade will harm MAP.)

2. There are other lines of evidence that catecholamine excess might be harmful too, like in certain sub-groups of CHF or stress cardiomyopathy--but it's always dangerous to draw pathophysiological analogies, and I appreciate that.

3. Patient-centred outcomes take a long time to adequately power, especially with sick patients, so isn't it fair to become responsive to meaningful physiological signals of benefit or harm (like with GFR)? It's like with NICE-SUGAR and the Leuvan trials. (Which I understand also burned early adopters and rightfully suggests caution here--but that was a long seven years!).

4. Still, fair points. In your view, is there at least equipoise?

This is really helpful for me as a student, and it's sometimes fraught to discuss with your own attending. But I don't want to overstep either. Feel free to shoot down as appropriate.

Well CO = SV X HR. In septic patients, if you treated them correctly you have maxed out their preload. If you decrease heart rate you are going to decrease CO by definition.

The main issue with taking these theoretical concepts and trying to apply them to patients is the potential for harm. I could give you a long, long, long list of things that should have made sense from a pathophysiological aspect but ended up harming people after they were studied more. Your upside is maybe a little GFR...the downside is rapid death in certain patients...

It all comes back to "do no harm." Probably, shouldn't just randomly try treatments with a significant chance of harm with very little data...
 
  • Like
Reactions: 1 user
Well CO = SV X HR. In septic patients, if you treated them correctly you have maxed out their preload. If you decrease heart rate you are going to decrease CO by definition.

It all comes back to "do no harm." Probably, shouldn't just randomly try treatments with a significant chance of harm with very little data...

What? Not at all.
1) SV is more than just preload.
2) CO = SV X HR. So drop HR but improve SV and you maintain CO (as presumably happened in this paper).
3) V-A coupling might be a better measure of cardiac efficiency than CO, EF, or SV anyway.
4) Still, doesn't matter. Perfusion and cardiac efficiency matter. Measures are measures, with limitations and shortcomings for all of them.
5) Reasons in favour of improved perfusion and cardiac efficiency with esmolol: SV includes diastolic ventricular filling, cardiac perfusion occurs in diastole, MAP is more than just cardiac efficiency in the form of CO, and improved arterial elastance--all of which could explain the drop in pressor support.

Surely there's at least equipoise: the physiological rationale is there for further investigation. In the thread from 2012, people seemed to be very strongly against the concept to begin with.

You raise an interesting point about preload, and the winds seem to be shifting back toward running patients drier. I don't know how that ties into beta-blockade, since all of the patients studied thus far were "drowning" in preload (if you want to use Marik's phrase). Beta-blockade might be much less helpful on drier patients (if it is indeed helpful at all).

"Not ready for prime-time" as said above. Fair enough.
 
Last edited:
What? Not at all.
1) SV is more than just preload.
2) CO = SV X HR. So drop HR but improve SV and you maintain CO (as presumably happened in this paper).

Okay you're flipping back and forth between different concepts. What are you talking about treating with b-blockers? I've treated tachycardia in septic patients with esmolol, but that is when they were chugging along with their baseline HR of 120 and then next thing we know the rate is 190 and BP has dropped out. With that, I am treating a pathological tachycardia which is directly impeding diastolic filling.

Your septic patient with a hr of 120s is not impairing their diastolic fill due to that heart rate. The decrease in HR you cause is not likely to provide you with a simultaneous increase in SV enough to maintain your CO, and the patient will crash.

If you're talking about using b-blockers to decrease overall adrenergic stimulation, improve vascular tone blah blah that's fine, but just realize that you're talking about two different indications for your b blocker therapy.
 
Members don't see this ad :)
Your septic patient with a hr of 120s is not impairing their diastolic fill due to that heart rate. The decrease in HR you cause is not likely to provide you with a simultaneous increase in SV enough to maintain your CO, and the patient will crash.

Friend, a major point of the paper I posted (plus the upcoming ESMOSEPSIS study and a few others) was to argue that this very idea is wrong (with the caveats about haemodynamic optimisation with fluids, pressors, HR-directed therapy, etc.)
 
Last edited:
Okay you're flipping back and forth between different concepts. What are you talking about treating with b-blockers? I've treated tachycardia in septic patients with esmolol, but that is when they were chugging along with their baseline HR of 120 and then next thing we know the rate is 190 and BP has dropped out. With that, I am treating a pathological tachycardia which is directly impeding diastolic filling.

Your septic patient with a hr of 120s is not impairing their diastolic fill due to that heart rate. The decrease in HR you cause is not likely to provide you with a simultaneous increase in SV enough to maintain your CO, and the patient will crash.

If you're talking about using b-blockers to decrease overall adrenergic stimulation, improve vascular tone blah blah that's fine, but just realize that you're talking about two different indications for your b blocker therapy.

And what concepts, exactly, am I flipping back and forth? Arterial elastance and cardiac efficiency are not inextricable. You can't just say the heart is the heart, and the vascular tree is something else entirely--even with relatively pure beta-1 blockade, and especially in a pathophysiological state like sepsis. That's the whole point of ventriculo-arterial coupling as a concept (see the article below).

The one consistent finding among all the studies thus far (not saying this is generalisable) is that pressor requirements drop with esmolol in hyperkinetic sepsis after haemodynamic optimisation with fluids and pressors.

http://ccforum.biomedcentral.com/articles/10.1186/cc12522
 
Last edited:
Your septic patient with a hr of 120s is not impairing their diastolic fill due to that heart rate. The decrease in HR you cause is not likely to provide you with a simultaneous increase in SV enough to maintain your CO, and the patient will crash.

Just a slight modification here. A heart rate of 120 could impair stroke volume if the LV or RV have diastolic dysfunction or significant hypertrophy. A rate of 120 per se is unlikely to be detrimental to a healthy rate, but someone with significantly impaired ventricle, it could be, though I would suspect that HRs in the 140s to 150s would be more detrimental. That all being said, a healthy or relatively healthy heart is not really effected by sinus tachycardia at the rates we are talking about (case in point, septic children under maybe 10 or so will typically have sustained HRs in the 170s to 180s with sepsis). Typically, sepsis as a clinical syndrome is unlikely to make a healthy heart into a sick heart. Though animal models may suggest that sepsis itself can be directly injurious to the myocardium, in clinical practice, septic patients don't usually die from heart failure.

Either way, attempting to control a normal, physiologic response to illness just to make the number better seems far fetched to achieve any significant outcome (again, unless the patient has significant heart issues at baseline).
 
Last edited:
  • Like
Reactions: 1 users
Typically, sepsis as a clinical syndrome is unlikely to make a healthy heart into a sick heart. Though animal models may suggest that sepsis itself can be directly injurious to the myocardium, in clinical practice, septic patients don't usually die from heart failure.

Either way, attempting to control a normal, physiologic response to illness just to make the number better seems far fetched to achieve any significant outcome (again, unless the patient has significant heart issues at baseline).

Fair points. Good post. I still think there's at least equipoise for the reasons stated above. Wait and see.
 
Last edited:
Agree with much of the above.

On the other hand, one cannot not notice the bad outcomes associated with prolonged excessive tachycardia in some patients, so I think a very short-acting easily titratable beta1-antagonist is not a bad idea, when the risks of not giving it are higher than the risks of giving it. Also, as an anesthesiologist, everything I do is a therapeutic trial that needs frequent reassessment, so I would not shy away even from combinations like esmolol plus phenylephrine, and then monitor the patient (preferably with cardiac echo) to see if it works. We have almost a dogma about not giving beta blockers in sepsis, the same way we have one about avoiding phenylephrine.

IMO, CCM is not and should never become cookie-cutter medicine. Some patients will do well on one thing, some patients on other. The same way some respond to norepinephrine, while some thrive on epinephrine, there is a population of patients that will enjoy a break from the cardiac effects of circulating catecholamines. We just have to figure out which population (my bet is on the elderly and those with known heart disease).

The same way too much fever or inflammation etc. can be bad, I am sure too much tachycardia can be harmful, even if physiologically appropriate. We have been taught the dogma that one doesn't treat sinus tachycardia, but its causes, and I am not so sure that's always right (e.g. the beneficial effects of beta-blockers in chronic HF).

Kudos to @lymphocyte for being this interested in critical care. I would just be both less aggressive and less defensive when chatting with people with much more clinical experience. ;)
 
Last edited by a moderator:
  • Like
Reactions: 1 users
As a quick point - when you get to residency (or even M4 year) your primary job will be to listen to more senior people and enact their plans and contribute more as you become comfortable and more senior. If you start to debate literature and become aggressive with these people it will lead to some rough experiences, for better or worse. A lot of big egos in medicine (which helped them get so far), and clinical exposure and experience is every bit as important as staying up with the literature. But great discussion.
 
  • Like
Reactions: 1 user
I tried to hyperlink (fail) the new Morelli paper from 2016:

http://link.springer.com/article/10.1007/s00134-016-4351-2

I keep saying "Morelli" because it's not just single center, it's the same lead author and non-blinded, so I'm naturally skeptical. Very strongly agreed about the Rivers' trial and all the lessons to be learned from it.

One challenge I've noticed with intensive medicine is adequately powering trials, since each patient is so nuanced and tricky. That makes physiological arguments a little more attractive, like with this new paper. I was wondering if these arguments were convincing anybody. But "not quite yet" is the sense I'm getting and fairly so.

ESMOSEPSIS from France should be out next year, so I guess we'll have more data then. But it's still open-label, industry-sponsored, with the primary outcome being cardiac index, and the secondary outcomes being pressor requirements and measures of microcirculatory flow--not exactly the hardest of end-points. Cocchi is running the study at BID with a completion date of 2019. Either way, it's going to be a long wait till 1) we get useful mortality data, and 2) any kind of respectable power across the hyperkinetic septic spectrum.

I'm a bad doctor who hasn't been keeping up the literature as much as I'd like. This new study is news to me.

And while it is interesting . . . I'm not convinced a give any rips necessarily about the statistical crap they threw against the wall looking for something to stick. So unless I missed it . . . did it improve anyone's survival? Or any other important end point? If not, then . . . "neat" but get back to me when we are saving lives with this.
 
  • Like
Reactions: 1 user
I'm a bad doctor who hasn't been keeping up the literature as much as I'd like. This new study is news to me.

And while it is interesting . . . I'm not convinced a give any rips necessarily about the statistical crap they threw against the wall looking for something to stick. So unless I missed it . . . did it improve anyone's survival? Or any other important end point? If not, then . . . "neat" but get back to me when we are saving lives with this.

I strongly disagree with the first part of your first sentence (well, you never know, but still). Also:

1. The 2013 study did indeed show a mortality benefit, but the mortality benefit was too good for some--didn't pass the sniff test (among the other shortcomings as mentioned above). 28-day mortality was 49.4% in the esmolol group vs 80.5% in the control group. Really?? And 80.5% mortality in controls? Holy buckets Batman! When in Rome, don't get sepsis... or are least try and be in the active arm of an open-label trial...

2. This study registered its primary and secondary outcomes ahead of time, which makes me less inclined to believe in statistical jiggery-pokery.

3. The thesis of this study was that sinus tachy is not just a "normal response" to sepsis; it might also facilitate a pathological interaction between the heart and the arterial tree, i.e. V-A decoupling. Hence: a reduction in HR presumably improved arterial elastance (a big component of V-A coupling) while still allowing for adequate systemic perfusion (since CO was unchanged despite noradrenaline requirements dropping). I know, yadda yadda yadda. And that's fair. Most of the research just seems to be arguing for physiological equipoise so the mortality trials can begin.

4. Clearly, there are early and late adopters in intensive care. In this case, I'm thinking no adopters.
 
Last edited:
4. Clearly, there are early and late adopters in intensive care. In this case, I'm thinking no adopters.

Well, there is some good reason for not being an early adopter. There have been a number of studies, specifically in critical care, that have shown large survival benefits in the initial go around with subsequent, larger, multi-center studies showing no benefit. So there should be some healthy skepticism when it comes to adopting new treatments, especially if it seems too good to be true. I have seen people harm patients through early adoption of single interventions, tight-glycemic control specifically. If there wasn't some healthy skepticism, we all would be slamming blood products into septic patients to keep their hemocrit >30% and giving fluids till their CVP was 12 mm Hg as we estimate their cardiac output through a PA catheter, all the while, making sure we are keeping their blood glucose between 80 to 120 mg/dl and making sure they get activated protein C. These were all shown to have huge benefits in the initial studies, with subsequent studies showing none and in some cases, harm.
 
  • Like
Reactions: 1 user
It blows my mind that they would use a swan ganz catheter instead of transthoracic or trans esophageal echo. With the echo it would be easy to tell the EF, systolic or diastolic failure and than separate them into groups. If there is a benefit from this treatment (BIG IF), I would bet the diastolic failure group would benefit most.
 
  • Like
Reactions: 1 user
It blows my mind that they would use a swan ganz catheter instead of transthoracic or trans esophageal echo. With the echo it would be easy to tell the EF, systolic or diastolic failure and than separate them into groups. If there is a benefit from this treatment (BIG IF), I would bet the diastolic failure group would benefit most.

My understanding of this study was to establish physiological rationale, and thermodilution gets at some physiological parameters better. Care was consistent with Surviving Sepsis (take that for what you will).
 
Last edited:
It blows my mind that they would use a swan ganz catheter instead of transthoracic or trans esophageal echo. With the echo it would be easy to tell the EF, systolic or diastolic failure and than separate them into groups. If there is a benefit from this treatment (BIG IF), I would bet the diastolic failure group would benefit most.

You mean hallucinograms?
 
It blows my mind that they would use a swan ganz catheter instead of transthoracic or trans esophageal echo. With the echo it would be easy to tell the EF, systolic or diastolic failure and than separate them into groups. If there is a benefit from this treatment (BIG IF), I would bet the diastolic failure group would benefit most.

I don't get this. You want to do TTE on all the patients with heart issues admitted to the ICU for sepsis to risk stratify? Um, if they're septic or intubated, how does this complicate getting a TEE which is much more accurate? I say this because my understanding is that TTEs in a sick tachycardic heart are less accurate, but I can't say how the accuracy or by how much it is affected. For either, one would have to figure out it was still good enough data to use for risk stratifying.

Is all this really going to generate data that is at least equivalent or more useful, more feasible, more cost effective, and with less risk than a Swan Ganz? I'm just curious. I don't know but this all went through my mind because I just remember that TTEs at least are not great in sick tachy hearts which is why I got the smackdown on ordering them in that population. IIRC, some patients would get a beta blocker just to get a TTE that was meaningful.
 
making sure they get activated protein C. These were all shown to have huge benefits in the initial studies, with subsequent studies showing none and in some cases, harm.

I just noticed what you said about activated protein C. I think that might have been another casualty of Evidence-Based Medicine.

PROWESS-SHOCK looked at a broad population for severe sepsis. But each pathophysiological state is unique. APC might better work, for example, in meningococcal sepsis with DIC or certain other coagulopathic phenotypes, and several retrospective studies said it does. First principle reasoning says it does. But who cares? As soon as PROWESS-SHOCK came out, DrotAA was withdrawn from the market worldwide. So much for that idea.

Veldman A, Fischer D, Wong FY, et al. Human protein C concentrate in the treatment of purpura fulminans: a retrospective analysis of safety and outcome of 94 pediatric patients. Crit Care 2010;14:R156-R156

Vincent J-L, Nadel S, Kutsogiannis DJ, et al. Drotrecogin alfa (activated) in patients with severe sepsis presenting with purpura fulminans, meningitis, or meningococcal disease: a retrospective analysis of patients enrolled in recent clinical studies. Crit Care 2005;9:R331-R343
 
Last edited:
I don't get this. You want to do TTE on all the patients with heart issues admitted to the ICU for sepsis to risk stratify? Um, if they're septic or intubated, how does this complicate getting a TEE which is much more accurate? I say this because my understanding is that TTEs in a sick tachycardic heart are less accurate, but I can't say how the accuracy or by how much it is affected. For either, one would have to figure out it was still good enough data to use for risk stratifying.

Is all this really going to generate data that is at least equivalent or more useful, more feasible, more cost effective, and with less risk than a Swan Ganz? I'm just curious. I don't know but this all went through my mind because I just remember that TTEs at least are not great in sick tachy hearts which is why I got the smackdown on ordering them in that population. IIRC, some patients would get a beta blocker just to get a TTE that was meaningful.

I think it's more TOE (that's Australian for TEE) vs Ganz. TTE would not be appropriate here for the reasons you mention and others. And they actually used both (for different physiological parameters). All they really cared about was the physiology.

If you check out the paper, the arterial waveform picture is pretty compelling (though not at all any kind of convincing), and immediately gets the point across that this concept has little to do with protecting the heart but economising the interaction between the heart and the arterial tree.
 
Last edited:
This really isn't a patient centered outcome. Unless these patients end up with a less long-term CKD, lower mortality, or something else that matter who cares if you affect the GFR slightly for a small period of time.

This is an antidote, but the "cleanest kill" in medicine I saw involved a beta-blocker in sepsis. A nurse was concerned about a tachy septic patient on multiple pressors and asked for medication to "treat the heart rate." Intern gives some lopressor. 5 minutes later patient is pulseless and code blue is called. 15 minutes later time of death is called. Patients with crappy hearts need a high heart rate during sepsis to maintain blood pressure. Taking that away can be fatal.

GFR's not a bad proxy for perfusion. Otherwise ditto the above.

Chime in my own anec-data. The cleanest kill I saw - also a beta blocker in sepsis. A little different, there was more indication to use a beta blocker than what you describe, but still deadly.

Came ITT to check out lymphocyte's post, & offer what I always call my "idiot analysis." Basically, I come at things like I don't know anything, ask stupid questions, bring up stupid points. Can have benefit of shining light on things, because the idiot analysis is often overlooked by docs.

I agree with all of the excellent physiology arguments that are suggesting extreme caution with beta blockers on a case by case and avoiding them in sepsis where you can. I think in cases where you want to apply medicine in an almost counter-intuitive way, or against common sense, which is sort of what I think using a beta blocker in "patients with crappy hearts that need a high heart rate during sepsis to maintain blood pressure" and "taking that away can be fatal" and amounts to, you need good evidence. You're making a case that it makes physiological sense except it sorta doesn't to me.

You mentioned early vs late adopters. This is getting into some of the psychology of practicing medicine. There are interventions we do that take time to see affect, and can have small effect sizes, or even big ones but further down the pipeline. In those cases, I think there's enough disconnect between action and effect to more easily convince people with stats and papers to change practice, and the burden of evidence I expect is lower.

When you have docs that have used beta blockers in sepsis and watched a patient die in front of them within minutes, or a lot of the experience here where BP is low, HR is high, and pressors are up, it's just gonna be a no go until it's more convincing and it doesn't seem like a drop in HR = drop in BP.
 
Last edited:
the idiot analysis is often overlooked by docs.

Wait a second, I thought I was offering the idiot analysis in this thread. I certainly feel that way a lot.

But sometimes the idiot analysis is idiocy only because it cuts against the received wisdom. Whatever you think of Thomas Kuhn, he did have a point about the role of anomalous results in preempting the paradigm. That's what's happening here, if you take the data seriously. And speaking of paradigm, look at the consistent theme in the thread: it's the heart, the heart is not the problem, the heart is fine in sepsis--everything is being thought of just in terms of the heart. But what if beta-1 blockade affected the arterial system in ways other than pure pharmacology? Maybe the heart and vascular tree are not so inextricable (at least not in sepsis)... that's why the arterial wave-form analysis was so compelling to me.

Anyways. Yadda yadda yadda. It's a skeptical bunch, and rightfully so. We'll just have to wait and see for (more convincing) mortality data.
 
Last edited:
Wait a second, I thought I was offering the idiot analysis in this thread. I certainly feel that way a lot.

But sometimes the idiot analysis is idiocy only because it cuts against the received wisdom. The herd mentality is alive and well in medicine, especially when it comes to herd skepticism. Remember what they used to say about beta blockers in HF?

Of course, silly ideas are silly until they're not. That's how medicine takes big leaps forward instead of just fiddling on the margins. Whatever you think of Thomas Kuhn, he did have a point about the role of anomalous results in preempting the paradigm. That's what's happening here, if you take the data seriously. And speaking of paradigm, look at the consistent theme in the thread: it's the heart, the heart is not the problem, the heart is fine in sepsis--everything is being thought of just in terms of the heart. But what if beta-1 blockade affected the arterial system in ways other than pure pharmacology? Maybe the heart and vascular tree are not so inextricable (at least in sepsis)... that's why the arterial wave-form analysis was so compelling to me.

Anyways. Yadda yadda yadda. It's a skeptical bunch, and rightfully so. We'll just have to wait and see for (more convincing) mortality data.

I read all the papers you linked and I was much more impressed with all of the flaws found in the study and its measures to be appreciably swayed by what looks like a decent sounding theory with what looks like narrow application. You can't take this data seriously.

It's missing the forest for the trees anyway.

The problem is, if you're wrong, the patient dies. Fast. Or maybe they just get an MI, ischemic brain injury, or renal failure from hypotension. If they're really in septic shock they're already on death's door and the ability of their body to compensate is already severely compromised. One of the most basic ways they are compensating is their heart rate.

So you're basically gambling that whatever you lose in output from lowering heart rate is gonna be made up elsewhere or by this coupling. I think it's a helluva gamble given how dysfunctional and unpredictable the peripheral vascular system behaves in sepsis. I dunno, one definition I read of sepsis basically said you can think of it as a cytokine storm that causes vascular dysfunction. Maybe there is some role in calming the storm with a beta blocker. You just said in sepsis the heart is fine - I think it's just as likely you put the "brakes" on the only part of the system that's acting appropriately.

It would be sort of horrifying if the data found that if this was applied widely that the overall effect would be reduction in mortality rate. Instead of a greater number of patients taching along before a higher proportion succumbed later, we could have the trade off giving a beta blocker, being wrong, losing the gamble, getting a quick blue kill, but the ones that continue on with their comfortable resting HR in the 80s live. It's like throwing people off a boat before it can sink and drown the rest. Cull a few to save the herd. :grumpy:
 
I read all the papers you linked and I was much more impressed with all of the flaws found in the study. You can't take this data seriously.

It's missing the forest for the trees anyway.

1. What flaws did you find impressive? Even the neutral editorial piece, which I didn't link, thought the study was at least well-designed. Do you think there's at least equipoise? Because if not, there are 5 or so active trials that shouldn't have gotten ethics approval--and all of those intensivists seem to take the data seriously enough.

2. I don't quite agree with that definition of sepsis (which itself is a pathophysiologically heterogeneous entity, far beyond just clinical definitions offered by Surviving Sepsis or Sepsis 3 or whatever), and I'm not sure how esmolol at 24 hours in an HR-directed fashion after haemodynamic optimisation would necessarily result in any kind of "quick kill." The example above was for metoprolol.
 
Last edited:
1. What flaws did you find impressive? Even the neutral editorial piece, which I didn't link, thought the study was at least well-designed. Do you think there's at least equipoise? Because if not, there are 5 or so active trials that shouldn't have gotten ethics approval--and all of those intensivists seem to take the data seriously enough.

2. I've already been called-out for being too aggressive with clinicians, so I'll hold my tounge, but I don't quite agree with that definition of sepsis (which itself is a pathophysiologically heterogeneous entity, far beyond just clinical definitions offered by Surviving Sepsis or Sepsis 3 or whatever), and I'm not sure how esmolol at 24 hours in an HR-directed fashion after haemodynamic optimisation would necessarily result in any kind of "quick kill." The example above was for metoprolol.

http://lifeinthefastlane.com/ccc/sepsis-definitions/

all I'll say is to point out I said ONE definition, meaning it's one way of looking at it. Obviously it's more complicated than what I said.

I went back to read the articles you linked to try to pick out points - and basically, all of it.
http://link.springer.com/article/10.1007/s00134-016-4407-3
Is why I hate the idea. It seems like someone has fallen in love with the math and theory and has lost sight of being practical.
I don't want to sit here and summarize the whole article - if you push, I might.

The next point, is that even the article that you presented that was in favor of it, actually provided me with the point that you're putting brakes on a system that could be at the patient's peril.
"However, the therapeutic concept of “setting the brake” by β-blockade seems deceptively simple and could be effective in carefully chosen septic shock patients (excluding those with hypovolemia, known complex cardiac comorbidities, tachyarrhythmias, hemodynamic instability despite vasopressor treatment, or systolic cardiac dysfunction)."

I agree, deceptively simple, relies on a lot of math in the face of common sense, and the above rules out so many patients in the ICU, to essentially be useful in patients that are cardiovascularly healthy - oh yeah, except for being in septic shock and on pressors. That is a very narrow demographic of septic shock.

"So the important, conceptual cardiovascular question still needs to be answered: how much preload is necessary for safe and effective β-blockade in septic shock?"

Part of the challenge of preload in these patients and why we have to pour so much fluid in them is because they 3rd space it - even with perfect hearts, because of the peripheral vascular dysfunction I mentioned. And as you mentioned, the heterogeneity is such that you don't know how much fluid it will take until you start pouring it in, and it can be a challenge not to overshoot. So this whole theory rests on fluid balance which is one of the most important things we manage in the hospital - and a huge PITA. It's a moving target that can only be assessed clinically. You can't go by numbers you have to listen to the lungs and squeeze the ankles.

So you're relying on what the fluid level is in a leaky boat for applying this brake.

As far as crumping the patient with esmolol vs metoprolol - fair point. I looked this up, but a 1/2 life of 9 minutes? I guess if you're following the above exclusions I'm less worried that 20 min (let's say) of decreased CO from esmolol is going to be unsurvivable.

On the other hand, 20 minutes is all it takes to watch someone lose consciousness, turn blue, and die. I'm seen it more than once from HR and BP drop. So I don't know that the the half life of esmolol vs metoprolol on the brake makes me feel any safer using it.

But I don't claim to have a lot of experience with esmolol and admittedly the n=2 quick clean kills described in this thread were metoprolol. These are my thoughts based on the reading and what I've seen. Admittedly, and maybe it's telling, I probably have PTSD from witnessing a beta blocker induced death in a septic patient.

To balance all these "caveats" in the positive review article, they go one about the immune system, and gene expression, and etc etc. That's what I meant by someone fell in love with theory and missed the forest for the trees. None of that **** matters if the patient isn't perfusing their limbs because I dropped their HR and CO.
 
"However, the therapeutic concept of “setting the brake” by β-blockade seems deceptively simple and could be effective in carefully chosen septic shock patients (excluding those with hypovolemia, known complex cardiac comorbidities, tachyarrhythmias, hemodynamic instability despite vasopressor treatment, or systolic cardiac dysfunction)."

It's deceptively simple because it probably has little to do with the heart itself except in so far as the heart might be a contraindication (like septic cardiomyopathy) or a further indication (like systolic cardiac dysfunction). V-A coupling is at issue here. The original study was merely one take on the "why" of 7 published or soon to be published trials on beta blockers in sepsis.

The main point of the "No" article was that V-A coupling was wrongly measured or a poor measure of cardiac efficiency to begin with, but even then: "the theoretical background is firmly laid." (And MAP minus Pdic also improved, which argues in favour of esmolol but for different reasons.) I didn't see that Prof McLean wrote the piece, but I know him and can ask about any specific points you (or anybody else) might have.

Anyways. I don't want to argue too much in favour of this idea. I keep saying it, but I'll say it again: equipoise. This is not all a condemned idea and doesn't belong in the graveyard yet.
 
Last edited:
I don't get this. You want to do TTE on all the patients with heart issues admitted to the ICU for sepsis to risk stratify? Um, if they're septic or intubated, how does this complicate getting a TEE which is much more accurate? I say this because my understanding is that TTEs in a sick tachycardic heart are less accurate, but I can't say how the accuracy or by how much it is affected. For either, one would have to figure out it was still good enough data to use for risk stratifying.

Is all this really going to generate data that is at least equivalent or more useful, more feasible, more cost effective, and with less risk than a Swan Ganz? I'm just curious. I don't know but this all went through my mind because I just remember that TTEs at least are not great in sick tachy hearts which is why I got the smackdown on ordering them in that population. IIRC, some patients would get a beta blocker just to get a TTE that was meaningful.
Theyre as good as the operator performing them...?

Sent from my SM-N910P using SDN mobile
 
  • Like
Reactions: 1 user
This really isn't a patient centered outcome. Unless these patients end up with a less long-term CKD, lower mortality, or something else that matter who cares if you affect the GFR slightly for a small period of time.

That's a fair point. I just want to say that Josh Farkas at PulmCrit has been advancing some very interesting arguments about the importance of renoresucitation in sepsis. And to be fair, Josh Farkas also proposed the "Bull**** Index," which I think I should take stock of with my long posts.

upload_2016-8-17_11-35-46.png




http://emcrit.org/pulmcrit/vanish-renoresuscitation-vasopressin-vepinephrine/#comment-262201
 
Last edited:
Top