The End of Early Goal Directed Therapy?

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

G-Man82

Full Member
15+ Year Member
Joined
May 16, 2005
Messages
401
Reaction score
190
So to the Critical Care guys here, now that the ProMISe Trial (as well as the Process and Arise trials) has finally been published, what's happening at your institutions? Has it caused any major changes in the way Sepsis is managed? I realize in the surgical setting, most patients already have lines, but are you routinely drawing ScvO2s still? Or transfusing?

The Surviving Sepsis Campaign will likely update their stance to take into account the current evidence. They say right now that since there is no harm in the use of EGDT, the bundles can still be used, until their latest revision is published. But it looks like EGDT may have fallen by the wayside, as long as fluids and antibiotics are started early.

The hospitals that I trained at had pocket-cards with the old Rivers protocol on them. They've since been destroyed and I think our computer systems managers are in the process of updating the sepsis/shock order sets that we have.

Just curious to see the response. Plus, I deliberately posted here, rather than the CC Forum. Those guys are welcome to post, too, but I was curious to hear the opinions of other Anesthesiologists, both general, CCM, or interest in CCM.

Members don't see this ad.
 
Last edited:
No changes made on our part. We have always followed a modified EGDT that made since to us. We would give moderate amounts of fluid, start levophed early, didn't transfuse until hemoglobin was bellow 7, never used dobutamine, never used a cvp, but instead did an echo to figure out the issue. Also we trend lactate but are ok with a slow correction in certain situations. Used concentrated albumin instead of crystaloid in some situations like low EF. Of course early cultures and antibiotics with double gram negative coverage and quickly de-escalate once cultures come back, source control, maybe check an scvo2. Our MICU colleagues on the other hand are lead by a guy who worked directly with Rivers at Henry ford before coming over and they more or less follow EGDT like the bible and flood their patients with a ton of fluid. To the point where surgery had to come and preform bedside laparotomies for abd compartment syndrome.
 
  • Like
Reactions: 2 users
I'm not a CCM guy, but I visit and chat with the team rather frequently. Our intensivists here are all medical, and rather inbred, so most practice in a very similar fashion. I can't remember the last time they regularly checked ScvO2 on septic patients (maybe an occasional one for curiosity), blood isn't transfused to arbitrary numbers, most will switch to albumin at some point during volume resuscitation, CVPs are largely ignored (SVV/PPV or TTE used for assessment of volume status), abx broad an early, dobutamine is very rare (generally just low-EF patients who are also septic).
 
Members don't see this ad :)
RJ, interesting that your MICU does that. Mine, and I just rotated through, is sometimes too restrictive with fluids, but they're very very willing to turn the Norepinephrine up to >100 mcg/min. I don't believe there's truly a max dose for a catecholamine, but at that dose, I think death is knocking. I'm not really a fan of albumin; I think it's just an expensive form of NS or LR, and I don't think it cuts down on total volume used by as much as we think it does. But nothing wrong with it.

Any of you guys using those Non-invasive CO monitors that also give you SVV, SVR, CO/CI, etc?
 
Yeah, our MICU will typically give 5 L of fluid or more. And their attendings are happy when they do. My attending explained it by saying that the chairman worked very closely with Rivers and they all drank the EGDT kool-aid.

In residency we used the vigileo and it worked ok. In fellowship we just do bedside TTE's, and just recently got hTEE's which is pretty cool but pricey. And I wouldn't exactly call them non -invasive.
 
Last edited:
  • Like
Reactions: 1 user
Think about the principles of EGDT and consider that there are now improved surrogates. ABX; then fluid resuscitate to a targeted endpoint (cvp is one, but a not very good one; try passive leg raise or improvement in PP or SV with fluid challenges; the pressors if MAP is still low; then a surrogate for cardiac function if lactate is not improving (SCVO2, cardiac index, etc...)
 
G man...we use the flotrac quite a bit, i like it when i get perforated bowel exlap septic guy to help determine fluid responsiveness , have used the tee probe in the past if we are truly struggling.

Anybody using ivc ultrasound?
 
1. Since the publication of EGDT the mortality for Sepsis has decreased tremendously. The original Rivers studies has baseline mortality in the 40% range. In the most recent studies the mortalities were in the high teens, which makes finding a mortality difference that is statistically significant harder to detect.

2. These new studies were comparing "usual treatment" vs EGDT. Many aspects of EGDT were readily accepted (early ABX, Fluid boluses early, ,pressers ) but many such as ScVo2, dobutamine, and transfusion threshold were not. So the fact that the new studies did not show the benefit of it once did is more a symbol of how Manny Rivers changed practice patterns over the last decade than anything else.

3. Our understanding of fluid management is being challenged all the time. In colorectal surgery we see that less aggressive volume resuscitation has better outcomes. In the ICU setting we all know that overaggressive resuscitation can lead to bowel dysfunction, pulm edema, cardiac dysfunction, kidney dysfunction from swollen glomeruli and the toxic nature of salt solutions. Resuscitation should have a dynamic goal that is used to know when enough is enough before you pass into the detrimental too much fluid scenario.

I see too many anesthesiologist focus on fluids when the SPV is flat and too many other Docs who keep giving fluid when the patient is suffering from ATN, Flooding an engine with gas when the injectors are clogged doesn't make it produce more horsepower!

IN the end, these articles have given credence to my practice of judicious fluid resuscitation, restricted transfusion strategies, and early broad ABX. Now New York state will have to pull back on the its strict recommendation to follow the EGDT for sepsis and the monetary penalties associated with non-compliance.
 
  • Like
Reactions: 2 users
It's interesting that no one has mentioned bedside ultrasound to eval IVC. I know it's not an ideal measure, but it's a piece of the puzzle.
 
A warning to those using the Flotrac- it works fine to trend changes in cardiac output related to volume challenges. It is a RANDOM NUMBER GENERATOR if the patient is on pressors.

I used to use it all the time, now I use it never. It can and will mislead you into thinking the patient has an adequate cardiac output when they don't if the patient is on pressors.

Echo is your best friend for determining volume status and appropriate fluid/pressor/inotrope therapy.
 
  • Like
Reactions: 1 user
1. Since the publication of EGDT the mortality for Sepsis has decreased tremendously. The original Rivers studies has baseline mortality in the 40% range. In the most recent studies the mortalities were in the high teens, which makes finding a mortality difference that is statistically significant harder to detect.

2. These new studies were comparing "usual treatment" vs EGDT. Many aspects of EGDT were readily accepted (early ABX, Fluid boluses early, ,pressers ) but many such as ScVo2, dobutamine, and transfusion threshold were not. So the fact that the new studies did not show the benefit of it once did is more a symbol of how Manny Rivers changed practice patterns over the last decade than anything else.

3. Our understanding of fluid management is being challenged all the time. In colorectal surgery we see that less aggressive volume resuscitation has better outcomes. In the ICU setting we all know that overaggressive resuscitation can lead to bowel dysfunction, pulm edema, cardiac dysfunction, kidney dysfunction from swollen glomeruli and the toxic nature of salt solutions. Resuscitation should have a dynamic goal that is used to know when enough is enough before you pass into the detrimental too much fluid scenario.

I see too many anesthesiologist focus on fluids when the SPV is flat and too many other Docs who keep giving fluid when the patient is suffering from ATN, Flooding an engine with gas when the injectors are clogged doesn't make it produce more horsepower!

IN the end, these articles have given credence to my practice of judicious fluid resuscitation, restricted transfusion strategies, and early broad ABX. Now New York state will have to pull back on the its strict recommendation to follow the EGDT for sepsis and the monetary penalties associated with non-compliance.
I know this thread was intended for anesthesiologists as noted by the OP, but I thought I'd chime in. I'm coming from the pulmonary/ccm side of things. I never understood the fascination with Rivers' study, and I still can't believe the SCCM was willing to base its entire Surviving Sepsis Campaign Guidelines off of a single center trial without subsequent validation. They certainly didn't rush to endorse widespread use of esmolol therapy after the Italian study was published in JAMA in 2013.

The overall in-hospital mortality in sepsis patients had been following for more than 20 years BEFORE the EGDT trial was completed. Martin, et al, published an interesting retrospective review of over 750,000,000 U.S. hospitalizations from 1979-2000, which included more than 10 million sepsis cases (Martin GS, et al. N Engl J Med 2003;348:1546-54). They have a nice figure that shows the overall mortality in this patient population steadily decreased in a linear fashion during that time period. I suspect sepsis-related mortality would have reached the high teens even without EGDT. That isn't to say that the trial was completely worthless. Certainly it brought increased emphasis on early recognition and antibiotic administration, which are clearly beneficial.

The medical intensivists in our facility primarily work in the MICU, SICU, and Burn ICU, and we largely abandoned the EGDT algorithm in mid-2013. Assessing fluid responsiveness is a major focus in our ICUs, so we tend to practice judicious fluid resuscitation as Seinfeld does (Burn ICU excluded). We have been using a combination of the NICOM, PiCCO, and US/TTE. Some of our surgeons like to use the VolumeView by Edwards Lifesciences. We use those monitors in conjunction with a small crystalloid bolus (250-500mL) or PLR maneuver when deciding whether to continue fluid resuscitation. That approach has certainly reduced the amount of "de-resuscitation" we've had to do.
 

Attachments

  • Epidemiology of Sepsis in U.S. from 1979-2000 - NEJM (2003).pdf
    118.5 KB · Views: 26
Last edited:
  • Like
Reactions: 1 user
I know this thread was intended for anesthesiologists as noted by the OP, but I thought I'd chime in. I'm coming from the pulmonary/ccm side of things. I never understood the fascination with Rivers' study, and I still can't believe the SCCM was willing to base its entire Surviving Sepsis Campaign Guidelines off of a single center trial without subsequent validation. They certainly didn't rush to endorse widespread use of esmolol therapy after the Italian study was published in JAMA in 2013.

The overall in-hospital mortality in sepsis patients had been following for more than 20 years BEFORE the EGDT trial was completed. Martin, et al, published an interesting retrospective review of over 750,000,000 U.S. hospitalizations from 1979-2000, which included more than 10 million sepsis cases (Martin GS, et al. N Engl J Med 2003;348:1546-54). They have a nice figure that shows the overall mortality in this patient population steadily decreased in a linear fashion during that time period. I suspect sepsis-related mortality would have reached the high teens even without EGDT. That isn't to say that the trial was completely worthless. Certainly it brought increased emphasis on early recognition and antibiotic administration, which are clearly beneficial.

The medical intensivists in our facility primarily work in the MICU, SICU, and Burn ICU, and we largely abandoned the EGDT algorithm in mid-2013. Assessing fluid responsiveness is a major focus in our ICUs, so we tend to practice judicious fluid resuscitation as Seinfeld does (Burn ICU excluded). We have been using a combination of the NICOM, PiCCO, and US/TTE. Some of our surgeons like to use the VolumeView by Edwards Lifesciences. We use those monitors in conjunction with a small crystalloid bolus (250-500mL) or PLR maneuver when deciding whether to continue fluid resuscitation. That approach has certainly reduced the amount of "de-resuscitation" we've had to do.

Maybe this is a dumb question, but are fluid responsiveness and being under resuscitated necessarily the same thing? I feel like a patient could be intravascularly replete, but not necessarily have their BP improve after IVF. Am I wrong?
 
A warning to those using the Flotrac- it works fine to trend changes in cardiac output related to volume challenges. It is a RANDOM NUMBER GENERATOR if the patient is on pressors.

I used to use it all the time, now I use it never. It can and will mislead you into thinking the patient has an adequate cardiac output when they don't if the patient is on pressors.

Echo is your best friend for determining volume status and appropriate fluid/pressor/inotrope therapy.

Agree with you wholeheartedly. I ended up doing some reading about those machines. Apparently they use a proprietary algorithm, which I think isn't revealed in it's entirety, (bis anyone?). So I'd hesitate to rely on it. If I have a question, TTE looking at the ivc is my go-to modality these days. Plus, in the era of low tidal volume ventilation, the number on those machines is pretty meaningless unless you temporarily increase the tidal volumes to roughly 10 ml/kg IBW, which people don't always do. If I look at anything on the monitors, then I usually look at PPV (adjusting for ventilation) on the arterial line waveform. I think that has been validated.

And the pulm/CCM view is welcome; I posted here to engage my fellow anesthesiologists.

But I believe a lot of what Rivers set out to prove did become part of usual care, hence the findings of no difference. The modifications are that I think transfusing to a number, or starting dobutamine, or just pounding someone with NS until they develop an abdominal compartment syndrome doesn't usually happen.

I read that JAMA article that was pointed out; in fact it came up in one of out CCM journal clubs. I'm not convinced.
 
Members don't see this ad :)
Maybe this is a dumb question, but are fluid responsiveness and being under resuscitated necessarily the same thing? I feel like a patient could be intravascularly replete, but not necessarily have their BP improve after IVF. Am I wrong?
That's a good question. The short answer is no, being fluid responsive and under-resuscitated are not necessarily the same thing. Fluid responsiveness is defined by an increase (often 10-15% depending on the research study) in SV/CO following administration of a fluid challenge (i.e. "auto-bolus" during PLR, IVF bolus, etc.). Being fluid responsive doesn't necessarily mean that one needs fluid. For instance, most normal humans walking around are in a state of fluid responsiveness. If you administered an IVF bolus to an average person walking down the street, his/her SV would likely increase. This has been studied in healthy volunteers. We certainly don't think most humans are under-resuscitated at baseline. Being fluid responsive implies that one is operating on the steep portion of the Frank-Starling curve.

The concept of fluid responsiveness is useful when deciding how to resuscitate/support your critically ill patient. If you find that he/she is not fluid responsive, then they are unlikely to benefit from additional IVF resuscitation at that point in time. If that particular patient is hypotensive and in shock, you would want to start some form of vasopressor support rather than continuing IVF resuscitation. Only about 1/2 of hemodynamically unstable critically ill patients are fluid-responsive (Marik PE, et al. Crit Care Med 2009, 37:2642-2647). The idea that every septic shock patient needs their "tank filled up" before you start vasopressor therapy is somewhat ridiculous. In general, septic shock is not a volume depleted state, and fluid therapy alone will not fix distributive physiology in most circumstances. I find myself agreeing with Paul Marik more often than not. He's definitely controversial at times, but his research and logic are sound. At our institution, we tend to use a similar approach to the one that he recommends. Fortunately, our ED physicians are on board with this approach. They don't necessarily assess for fluid responsiveness at the time of presentation, but they at least limit their empiric IVF boluses to no more than 20-30mL/kg up front. If the patient is still hypotensive at that point, then they start vasopressor therapy. Once they're in the ICU, we use some form of monitor (NICOM, PiCCO, etc) to help assess for further fluid responsiveness. We recheck this periodically (depends on the individual patient and circumstances). I've attached one of Marik's articles, if you're interested in reading more about the topic. You may have already read it.
 

Attachments

  • Hemodynamic Parameters to Guide Fluid Therapy - Ann Intensive Care (2011).pdf
    390.1 KB · Views: 34
Last edited:
But I believe a lot of what Rivers set out to prove did become part of usual care, hence the findings of no difference. The modifications are that I think transfusing to a number, or starting dobutamine, or just pounding someone with NS until they develop an abdominal compartment syndrome doesn't usually happen.

That may be true, but I still think Rivers gets too much credit. As I mentioned in my previous post, sepsis-related mortality has been falling steadily in a linear fashion since the late 1970s. If you read Martin's paper from NEJM 2003, you'll see that average sepsis-related mortality decreased from ~30% around 1980 to just below 20% around 2000. Now, they included all-comers with sepsis, and Rivers' study population was clearly on the sicker end of the spectrum. EGDT wasn't really implicated on a large scale until 2003/2004, and there really hasn't been much change in the mortality rate since that time. We're still hovering in the high teens to low 20s.

I read that JAMA article that was pointed out; in fact it came up in one of out CCM journal clubs. I'm not convinced.
I completely agree with this. The mortality rate in the control group was 80%!!! The patients were definitely sick. The average SAPS II score was in the low-to-mid 50s, which is typically associated with a mortality risk of ~50%. I don't think it's difficult show a mortality difference with any intervention in the ICU, if 4/5 patients in the control group die! Having said that, I'm intrigued by the idea of providing rate control in patients with significant diastolic dysfunction, but we certainly need much more robust studies before I'm willing to consider this on a broader scale.
 
  • Like
Reactions: 1 user
It's interesting that no one has mentioned bedside ultrasound to eval IVC. I know it's not an ideal measure, but it's a piece of the puzzle.

And since you brought it up, please share your interpretation of "IVC evaluation" with us in making diagnoses and choosing cardiovascular therapies in ICU patients.
 
A warning to those using the Flotrac- it works fine to trend changes in cardiac output related to volume challenges. It is a RANDOM NUMBER GENERATOR if the patient is on pressors.

Agreed that too little, or nothing, is known about how vasopressors affect pulse pressure variation, pulse contour analysis, or any other technologies built on such. My hunch would be that pressors artificially lower PPV by forcing previously "pooled" circulating volume into the central circulation. Does that mean that a PPV of 8 while on pressors is really high and fluid-responsive? What dose of pressors? If anyone knows, I'm listening.

I don't think I'll miss the days of a massive hyperchloremic acidosis from 5+ liters of NS being viewed as "normal," especially by medicine peeps :)
 
And since you brought it up, please share your interpretation of "IVC evaluation" with us in making diagnoses and choosing cardiovascular therapies in ICU patients.

Well, I should start by saying I'm an EM resident, not an intensivist (although I hope to be one eventually). Where I'm doing residency, it's rare for a patient to physically make it to an ICU bed within 6 hours, so EGDT or whatever variant of it is being practiced essentially falls entirely on our shoulders. I'll use bedside US on any patient I'm going to start pressors on. If I have a normal-sized adult that comes in for what seems to be septic shock (and no history of bad CHF or ESRD), I'll give them 2-3 L of fluid immediately (our RNs are great at getting 2 large bore IVs and cranking in the fluid when we ask them to). When it's time to start pressors, I'll grab a CVL kit and the ultrasound. I'll check the IVC before placing a CVL since I already had to get the machine. If the IVC is tiny and the heart is hyper-dynamic, I'll start levo and hang another 1-2 liters and re-check the IVC.

I realize that IVC isn't a great marker, but if it's flat the cardiac output approaches 100%, I feel comfortable aggressively hydrating.
 
That's a good question. The short answer is no, being fluid responsive and under-resuscitated are not necessarily the same thing. Fluid responsiveness is defined by an increase (often 10-15% depending on the research study) in SV/CO following administration of a fluid challenge (i.e. "auto-bolus" during PLR, IVF bolus, etc.). Being fluid responsive doesn't necessarily mean that one needs fluid. For instance, most normal humans walking around are in a state of fluid responsiveness. If you administered an IVF bolus to an average person walking down the street, his/her SV would likely increase. This has been studied in healthy volunteers. We certainly don't think most humans are under-resuscitated at baseline. Being fluid responsive implies that one is operating on the steep portion of the Frank-Starling curve.

The concept of fluid responsiveness is useful when deciding how to resuscitate/support your critically ill patient. If you find that he/she is not fluid responsive, then they are unlikely to benefit from additional IVF resuscitation at that point in time. If that particular patient is hypotensive and in shock, you would want to start some form of vasopressor support rather than continuing IVF resuscitation. Only about 1/2 of hemodynamically unstable critically ill patients are fluid-responsive (Marik PE, et al. Crit Care Med 2009, 37:2642-2647). The idea that every septic shock patient needs their "tank filled up" before you start vasopressor therapy is somewhat ridiculous. In general, septic shock is not a volume depleted state, and fluid therapy alone will not fix distributive physiology in most circumstances. I find myself agreeing with Paul Marik more often than not. He's definitely controversial at times, but his research and logic are sound. At our institution, we tend to use a similar approach to the one that he recommends. Fortunately, our ED physicians are on board with this approach. They don't necessarily assess for fluid responsiveness at the time of presentation, but they at least limit their empiric IVF boluses to no more than 20-30mL/kg up front. If the patient is still hypotensive at that point, then they start vasopressor therapy. Once they're in the ICU, we use some form of monitor (NICOM, PiCCO, etc) to help assess for further fluid responsiveness. We recheck this periodically (depends on the individual patient and circumstances). I've attached one of Marik's articles, if you're interested in reading more about the topic. You may have already read it.

Thanks!
 
Well, I should start by saying I'm an EM resident, not an intensivist (although I hope to be one eventually). Where I'm doing residency, it's rare for a patient to physically make it to an ICU bed within 6 hours, so EGDT or whatever variant of it is being practiced essentially falls entirely on our shoulders. I'll use bedside US on any patient I'm going to start pressors on. If I have a normal-sized adult that comes in for what seems to be septic shock (and no history of bad CHF or ESRD), I'll give them 2-3 L of fluid immediately (our RNs are great at getting 2 large bore IVs and cranking in the fluid when we ask them to). When it's time to start pressors, I'll grab a CVL kit and the ultrasound. I'll check the IVC before placing a CVL since I already had to get the machine. If the IVC is tiny and the heart is hyper-dynamic, I'll start levo and hang another 1-2 liters and re-check the IVC.

I realize that IVC isn't a great marker, but if it's flat the cardiac output approaches 100%, I feel comfortable aggressively hydrating.

OK so here's a question- what fluid are you using for these resuscitations? Please please please don't say NS.
 
And since you brought it up, please share your interpretation of "IVC evaluation" with us in making diagnoses and choosing cardiovascular therapies in ICU patients.
Haha just wanted to say this was almost word for word a question I was asked on my oral boards. After they heard my answer, for some reason they started asking me bunch of pain questions.
I'd like to contribute by saying fluid overload is a PITA. So stop doing it
 
  • Like
Reactions: 1 user
Then here are some links from your field. There are plenty from anesthesiology, but these may be more germane to your practice.

http://emcrit.org/podcasts/chloride-pressure-poisoning/

http://boringem.org/2013/01/06/normal-saline-the-coke-of-crystalloid-fluids/

Note that the cost of either Plasmalyte or Normosol, at least in my shop, is far lower than the $7/liter quoted in the above links, we get it for about $3/liter I believe.

I know, I know.....

The only isotonic fluids we stock in our pyxis are NS and LR. I would like to switch to a more balanced resuscitation fluid, but I work within the confines of my hospital.
 
Nothing wrong with LR.

Every c-section gets a few liters of LR. It's a perfectly reasonable resuscitation fluid, and far better than NS.
 
Top