Introduction with a case
Once upon a time at Genius General Hospital, an elderly man was admitted to the ICU with rapidly progressive cellulitis and tachypnea (1). We admitted him to the ICU despite a normal blood pressure because he looked toxic. Soon after arriving in the unit his blood pressure dropped, so vasopressors were started and lines were inserted.
A few hours later, I was called to his bedside due to refractory hypotension. He was on the maximal dose of three agents according to institutional guidelines (30 mcg/min norepinephrine, 10 mcg/min epinephrine, 0.04 U/min vasopressin), meanwhile phenylephrine was being up-titrated. His blood pressure was 60/30 on a femoral arterial line. It was suggested that he was dying and we should call his family in from the waiting room to say goodbye.
We told his family that things were looking dire, and they did come into the room for what we all anticipated was a last encounter. Meanwhile, we up-titrated his vasopressors to 80 mcg/min norepinephrine and 40 mcg/min epinephrine. His blood pressure steadily improved over the next 15 minutes. Within a few hours we were down-titrating the vasopressors. He eventually made a full recovery.
What is your hospital’s “maximum dose” of vasopressor?
Maximal doses used in clinical studies have ranged between 0.2-5 mcg/kg/min (Dopp-Zemel 2013). Every hospital and pharmacopeia have their own “maximum dose” of vasopressors. For example:
The maximum dose of vasopressors is important for a few reasons:
- Clinicians may be afraid to titrate above their hospital’s “maximum” dose, lest they run afoul of institutional policy.
- Failure of the patient to respond to “maximal” doses may be interpreted as meaning that the patient is moribund, with any further therapy being futile.
Theory: What is the maximum effective dose of vasopressors ??
Any drug ought to have a dose-response curve. Such curves indicate that above a certain dose, additional drug will have little effect. This would suggest that a maximal effective dose of vasopressor ought to exist. However, it also implies that higher doses of vasopressor aren’t dangerous (they are merely futile, since all the receptors at this point are fully saturated).
Dose-response curves don’t exist for humans due to ethical concerns. Dose-titration in rats shows that norepinephrine has a nearly linear effect up to a dose of 1.35 mcg/kg/min (Pang 1986). This suggests that commonly used clinical doses aren’t close to saturating all the alpha-receptors.
However, excessive norepinephrine doses could be dangerous within the context of an individual patient’s physiology. Specifically, excessive afterload could threaten to choke off cardiac output (thereby generating an iatrogenic state of vasopressor-induced shock). The norepinephrine dose at which this occurs would reflect several competing factors: cardiac function, volume status, and underlying vasoplagia. Thus, there might not exist any specific norepinephrine dose which would be expected to be detrimental for every patient. The same dose of norepinephrine that would kill one patient (e.g. someone with cardiogenic shock) might be required to keep another patient alive (e.g. someone with profound vasodilation).
Methodology pitfalls: Self-fulfilling prophecies & circular logic
Unfortunately, all available evidence consists of retrospective case series. Before considering the evidence, we must respect its limitations. One way to appreciate this is through the following thought experiments.
Thought experiment #1
- Imagine that there is a maximal effective dose of norepinephrine, let’s say 30 mcg/min. Any infusion above 30 mcg/min has exactly the same efficacy as 30 mcg/min.
- Imagine that we have a patient who is going to be hypotensive for one hour, and then her blood pressure will improve. This will happen regardless of vasopressor dose – she simply needs an hour to respond to resuscitation. Let’s imagine how this case could play out in two parallel universes:
- Universe A: Everyone in this universe believes that the maximal dose of norepinephrine is 30 mcg/min. Consequently, the norepinephrine will be titrated up to 30 mcg/min, but no higher. The patient will be hypotensive for one hour on 30 mcg/min norepinephrine, then her blood pressure will improve. Everyone in this universe concludes that the patient simply needed some time to recover.
- Universe B: Everyone in this universe believes that there is no maximal dose of norepinephrine. For the hour that the patient is hypotensive, norepinephrine is continually up-titrated. By the end of the hour, the norepinephrine has been increased to a dose of 200 mcg/min and the patient’s blood pressure finally improves. Everyone in this universe concludes that the high dose of norepinephrine caused her blood pressure to improve.
Thought experiment #2
In order to prove that high doses of norepinephrine are truly required, it would be ideal to perform ongoing,repeated dose-titration. For example:
- The patient appears to require 200 mcg/min norepinephrine to achieve a MAP >65mm.
- If the norepinephrine is decreased to 150 mcg/min this immediately causes hypotension, which immediately resolves when the norepinephrine is increased to 200 mcg/min. There is an obvious causal relationship between adjusting the norepinephrine dose and the blood pressure, which is reproduced several times.
In this scenario, it would be reasonable to conclude that the high dose of norepinephrine is actually needed. However, even in this situation it’s possible to fool yourself. For example, let’s perform the following thought experiment:
- Imagine that the maximal dose of norepinephrine is 30 mcg/min. Any higher dose has no additional effect.
- The patient’s blood pressure is oscillating between 55 mm and 75 mm. These oscillations are due to the patient’s intrinsic physiology; they have absolutely nothing to do with the vasopressor dose.
- The nurse is titrating the norepinephrine up and down between 100-200 mcg/min, based on the blood pressure.
- The resulting pattern will show that whenever the vasopressor dose is decreased, the blood pressure soon falls. Alternatively, then the vasopressor dose is increased, the blood pressure soon increases (graph above). This would seem to indicate that the vasopressor dose is driving the blood pressure. Of course, in this thought experiment we already know that this is purely a temporal correlation which is generated by the way the nurse is titrating the medication.
Thought experiment #3: The self-fulfilling prophecy
Imagine that some patients require very high doses in order to survive. Unfortunately, in one country it is incorrectly believed that the maximum dose of norepinephrine is 30 mcg/min. Any patient who fails to respond to 30 mcg/min is deemed to be moribund, leading to withdrawal of care. Consequently, no patient ever receives >30 mcg/min of norepinephrine. This creates a self-fulfilling prophesy that the maximum dose of norepinephrine is 30 mcg/min.
Clinical evidence
So, it’s possible for retrospective case reports to be easily confounded. That said, this is the best data that we have, so let’s take a look at it.
Auchet et al 2017
This is a retrospective single-ICU French study of septic patients requiring >1 ug/kg/min vasopressor between 2008-2013. 106 patients required this dose, making up 15% of all patient treated for septic shock. The most commonly used treatment was norepinephrine monotherapy. On average, high-dose vasopressor was required for 84 hours. 28-day mortality was 60%. Among survivors, the maximum rate was on average 2.3 ug/kg/min. The average norepinephrine dose was fairly predictive of death (area under the AUC curve 0.76). 6% of patients suffered digital or limb necrosis.
Martin et al. 2015
This is a retrospective single-ICU French study of septic shock patients admitted from 2009-2013. In-hospital mortality of all 324 septic shock patients was 48%, which is rather high. 84 patients (one quarter) received a maximal dose of norepinephrine >1 mcg/kg/min, of whom 90% died.
Sviri et al. 2014
This is a retrospective single-center Israeli study of patients receiving vasopressors in a medical ICU between 2008-2010. 166 patients who received norepinephrine or epinephrine were included, of whom 51 received high-dose vasopressors (defined as >40 mcg/min). In-hospital mortality among all patients receiving any vasopressor dose was extraordinarily high at 75%. Vasopressor dose was fairly predictive of death (AUC 0.779). High-dose vasopressor use was associated with a 90% in-hospital mortality.
Brown et al. 2013
This is a retrospective study involving five US hospitals between 2005-2010 describing patients requiring high-dose vasopressor (defined as >1 ug/kg/min norepinephrine or equivalent doses of other vasopressors)(2). 443 patients were included, of whom 241 had septic shock. 90-day mortality was high in the entire group (83%) as well as the subgroup with septic shock (80%). Digital or limb necrosis occurred in only 8% of surviving patients. Vasopressor dose correlated with increased mortality as shown above.
Dopp-Zemel et al. 2013
This is a retrospective single-ICU study from the Netherlands between 2007-2009 involving 113 patients treated with >0.9 mcg/kg/min norepinephrine. 28-day mortality was 66%. A norepinephrine dose above 2.22 mcg/kg/min was associated with 100% mortality, but this is statistically insignificant due to the very low number of patients treated with this dose (n=3).
Caution about using high-dose vasopressors
The above studies show a strong correlation between high-dose vasopressors and mortality. This is probably because the need for high-dose vasopressors correlates with greater disease severity, not because high-dose vasopressors cause mortality. Nonetheless, the need for high-dose vasopressors should never be taken lightly. This is generally a treatment of last resort. Whenever high-dose vasopressors are needed, meticulous evaluation is needed (ideally including echocardiography) with particular attention to the following questions:
- Is blood pressure truly that low? Consider placement of a femoral or axillary arterial catheter to transduce central arterial pressure.
- Is there a role for volume resuscitation or inotropic support? Some patients who respond poorly to norepinephrine may do better with epinephrine (see: epinephrine challenge).
- Is there occult right ventricular failure that could be treated (e.g. with pulmonary vasodilation)?
- Is there low cardiac output and impaired perfusion (e.g. mottling)? In that case, additional vasoconstriction may simply aggravate matters.
- Is there a failure of surgical source control or incorrect antibiotic selection?
- Is the patient on adequate adjunctive therapy (stress-dose steroid, possibly thiamine/ascorbate)?
- Is there a pH abnormality that merits correction (e.g. treatment of hyperchloremic metabolic acidosis with bicarbonate)?
- Is there autoPEEP or elevated intra-abdominal pressure which is impairing venous return?
- Is the patient on any medications which may be reducing the blood pressure (e.g. propofol, dexmedetomidine)?
- Is there clinically significant hypocalcemia?
- Published series suggest that it’s not uncommon for septic patients to require high-dose vasopressors (above ~1 mcg/kg/min norepinephrine).
- Vasopressor dose is moderately predictive of mortality. However, no particular vasopressor dose is 100% specific for death. For example, patients can survive despite requiring extremely high doses (e.g. >2 mcg/kg/min).
- The concept of a “maximum dose” of vasopressors should be discouraged, as this appears to be a myth. Limiting vasopressor dosage below an arbitrary rate could prevent successful resuscitation of the sickest patients.
- High-dose vasopressors are a treatment of last resort. In this situation, treatable causes of hypotension should be aggressively sought and corrected.
Related
- Epinephrine challenge (PulmCrit)
- Vasopressor basics (EMCrit)
- Norepinephrine (Deranged Physiology blog)
Acknowledgement: Thanks to Dr. Gilman Allen for thoughtful comments on this post.
Notes
- Someone will probably ask: how do you get septic from cellulitis? Basically two ways: either necrotizing fasciitis or Group A streptococcal infection with toxic shock syndrome (and often bacteremia). This was the latter. Toxic shock syndrome is actually fairly common with invasive group A strep infections (more on this here).
- One weakness of this study is that it considered 1 mcg of norepinephrine to be equivalent to 2.2 mcg of phenylephrine. In contrast, most modern studies seem to use a conversion whereby 1 mcg of norepinephrine is equivalent to 10 mcg of phenylephrine. A 1-2.2 conversion rate of phenylephrine could artificially inflate the “effective norepinephrine” dose.
- Pulmcrit wee: The cutoff razor - April 15, 2024
- PulmCrit Blogitorial – Use of ECGs for management of (sub)massive PE - March 24, 2024
- PulmCrit Wee: Propofol induced eyelid opening apraxia – the struggle is real - March 20, 2024
Where would you place an axillary arterial line? Upper arm/armpit or anterior chest wall? Does this carry a high risk of thrombosis or incompressible bleeding?
Ultrasound-guided axillary line ends up pretty close to the armpit. Place carefully w/ ultrasound using a femoral A-line kit (it’s deep). It is compressible in theory although have never needed to compress it. I believe that available evidence suggests that this is a safe site although there isn’t tons of evidence. IMHO the axillary site is under-utilized whereas the brachial site is over-utilized (I worry more about clots in brachial site b/c smaller artery). For more on a-lines see Scott Weingart’s podcast here: https://emcrit.org/emcrit/arterial-lines/
Great points Josh, clearly you’ve been frustrated by the old “He’s on max vasopressors, have you called the family yet?” statement we tend to get from staff and perplexed by where the 30mcg/min “max” even comes from. I agree about the calcium, we all see pts who seem to have large improvements in BP when we give calcium and not everywhere has the capability of getting rapid turnaround on ionized calciums and corrected generally ones aren’t accurate in the ICU so some times you just gotta try pushing some calcium But re: the pH not being optimal…. I’m not really… Read more »
Giving bicarb for lactic acidosis is worthless (potentially dangerous), no argument there.
There is some animal and human evidence that hyperchloremic metabolic acidosis impairs hemodynamics (see #5 here https://emcrit.org/pulmcrit/smart/). I think giving bicarb is a reasonable approach in this situation but can’t prove that it’s beneficial.
Great post as usual. I too have been frustrated by this issue. Another paper providing support for good outcomes in extremely high vasopressor use (up to 100 μg/min norepinephrine, 150 μg/min epinephrine, 250 μg/min phenylephrine) was here:
https://www.ncbi.nlm.nih.gov/pubmed/23642908
It was in the setting of calcium channel blocker overdose, so a selected population, but still illustrating that when you need higher doses, they work.
Thanks, great reference. This supports the concept that we aren’t coming close to saturating the alpha-receptors at conventional doses of norepinephrine.
Very nice points. I’d like to add in a few words in support of POCUS and a-lines here. I realize the post is about escalating to high doses of pressors, but sometimes the problem is that vasopressors aren’t the answer anyway. As everyone here already knows, blood pressure is generated by the product of your CO and your SVR plus your CVP. Furthermore, CO depends on multiple factors such as HR, preload, filling time, inotropy, lusitropy, RV/LV interactions (in series and parallel), etc. All of this information can be reliably gathered with POCUS and acted upon in a competent way… Read more »
Absolutely agree. Whenever you are cranking up the norepinephrine and it’s not having much impact you need to think really seriously about whether you’re using the wrong agent or wrong strategy entirely. I think 97% of the time, high-dose vasopressors are NOT the answer to your patient with severe hypotension. However, if you’ve exhausted other options (or while you are running through other options), it’s probably better to use high-dose vasopressors than leave the patient with a blood pressure that is incompatible with life.
Yes, Josh, thanks! Your response hit on exactly my point. I’ll admit that I sometimes I’ve been guilty of increasing my vasopressor dose, only to realize that the patient actually did need some fluids, which cognitively I’d discounted because the patient had already had a liter or two. This is probably where the straight leg raise or volume challenge acutally contributes to the shocked patient (rather than on the front end of the resuscitation–maybe I’m wrong), especially if the patient is intubated and the ventilator causes a little confusion in the POCUS exam due to ventilator induced RV afterload dilating… Read more »
how might the advent of angiotensin II for septic shock be changing the game? where does that fit in your algorithm?
I don’t think there is enough evidence available to justify the use of Ang-II at this juncture. The evidence on this is very mixed: https://emcrit.org/pulmcrit/angiotensin-ii/.
Hi Josh,
I read your great article here and then stumbled upon your EPI challeng article form a while back. Wondering what your thoughts on using Aortic (or carotid, or any other) VTI to ascertain the effectivenss of an EPI challenge?
Interesting idea. Honestly I mostly go based on traditional metrics like blood pressure, urine output, appearance, heart rate, and skin perfusion. Must admit that I’ve had issues with regards to reproducibility of VTI over time and it’s also pretty labor intensive so I’m not a huge fan.
Hi Josh, really fascinating post. Regarding previous comments about calcium… what are your thoughts on levosimendan?
I don’t know. It’s not available in the US so I haven’t paid much attention to it, sorry.
Very good post. I don’t believe in a max vasopressors dose. If the patient needs higher doses, why not give it to him? In our ICU we have already given for some patient methylene blue or GIK (glucose, insulin, and potassion) infusion to win the patient some extra time until a mechanical cause is reverted. It works perfectly.
I have a (perhaps stupid) question that I’m hoping people can reconcile for me. A phrase I hear commonly is ‘target a MAP of xxx’. When I hear this, I think of the following questions: 1. Isn’t organ blood flow (not perfusion pressure) the main concern when it comes to perfusion? If so, why do we target BP rather than try to optimise flow? Which organ benefits by targeting a specific mean perfusion pressure? 2. What is the evidence that vasopressor use improves organ blood flow in critically ill patients and how does this correlate with BP? In healthy patients,… Read more »
OK, so I’m a year late to this 🙂 Nice article. Re last post, we target pressure rather than perfusion…because we can. Ever since Hales measured blood pressure in the horse. And pulse. These things are intuitive. Even the so-called ‘dynamic indices’ of perfusion only tell you what’s happening in the macrocirculation, not in the tissue beds. And since different organs have different perfusion needs, methods designed to measure single organ perusion as a global parameter necessarily fall short. When you get to low Reynolds numbers in the microcirculation it’s the (driving) pressure head that matters. Vasopressin appears to predominantly… Read more »
Hi,
Quick question as I’m trying to do a presentation that HDV (>0.5mcg kg min, levophed) is OK and can be beneficial for a pt, etc., and I’m looking at the Yang et al study. I’m not finding in the Yang paper the NE dose of 1.35/mcg/kg/min before it plateaus. I’m sure it’s a calculation I’m missing as they write it in mol/mcg/kg/min x 10^11. I just need to be able to explain that before adding it to my ppt. How did you get 1.35mcg/kg/min? Thank you much!!
Can we administer in continuous infusion at the same time Norepinephrine and adrenaline?
Hello Dr Farkas, thank you for this insightful article. I was wondering how you go about vasopressor dosing the morbidly obese population? thank you:))