Why I don’t give vasopressors in sepsis

It’s become popular to use the term ‘vasopressors’ or just ‘pressors’ when noradrenaline/norepinephrine or even (in some places still) dopamine are given. I have resisted this trend and continue to use the term ‘vasoactive’ drugs, on the basis that the effects they produce (and that we may desire) are not limited to a pure alpha adrenergic effect on vascular tone, but they have effects on heart rate and contractility too (as well as preload through venous effects). If you don’t believe me about noradrenaline/norepinephrine, then check out one of my favourite critical care papers of all time: the CAT study.

There are of course real pressors out there – phenylephrine acts on alpha receptors, as does methoxamine. Metaraminol predominantly acts on alpha receptors but does also cause some release of noradrenaline/norepinephrine.

Why is this important? All these drugs will fix hypotension, right? Yes, they should. However should blood pressure be our main treatment goal? What we’re really interested in is organ perfusion, which depends on regional blood flow to vital organs. It’s possible that a drug could fix the measured blood pressure and give a nice ‘macroscopic’ number, while at the same time reducing cardiac output and adversely affecting regional blood flow to organs through local vasoconstrictive effects. My view is that this is more likely with pure ‘pressors’ (like phenylephrine), which is why I avoid them in septic shock and opt for catecholamine infusions (noradrenaline/norepinephrine).

This is important in my practice setting of retrieval medicine, where, prior to interfacility transport, physicians might sometimes be tempted to ‘push pressors’ peripherally rather than insert a central venous catheter and commence a catecholamine infusion. While the former approach might be more expeditious and make the vital signs chart look pretty, one wonders about what effect this is having on tissue oxygen delivery.

A fascinating review of papers on pressor physiology1 suggests these agents have the following effects:

  • conflicting data on changes in myocardial perfusion
  • increase both left and right heart afterload
  • decrease venous compliance with the potential to increase venous return although the impact of this on cardiac output is controversial
  • controversial effect on cerebral bloodflow
  • decrease bloodflow to the kidneys
  • adverse affects on gastrointestinal tract bloodflow

Phenylephrine and methoxamine are direct-acting, predominantly α(1) adrenergic receptor (AR) agonists. To better understand their physiologic effects, we screened 463 articles on the basis of PubMed searches of “methoxamine” and “phenylephrine” (limited to human, randomized studies published in English), as well as citations found therein. Relevant articles, as well as those discovered in the peer-review process, were incorporated into this review. Both methoxamine and phenylephrine increase cardiac afterload via several mechanisms, including increased vascular resistance, decreased vascular compliance, and disadvantageous alterations in the pressure waveforms produced by the pulsatile heart. Although pure α(1) agonists increase arterial blood pressure, neither animal nor human studies have ever shown pure α(1)-agonism to produce a favorable change in myocardial energetics because of the resultant increase in myocardial workload. Furthermore, the cost of increased blood pressure after pure α(1)-agonism is almost invariably decreased cardiac output, likely due to increases in venous resistance. The venous system contains α(1) ARs, and though stimulation of α(1) ARs decreases capacitance and may transiently increase venous return, this gain may be offset by changes in afterload, venous compliance, and venous resistance. Data on the effects of α(1) stimulation in the central nervous system show conflicting changes, while experimental animal data suggest that renal blood flow is reduced by α(1)-agonists, and both animal and human data suggest that gastrointestinal perfusion may be reduced by α(1) tone.

A review of clinical articles2 reveals few evidence-based indications for true pressors. Possible situations where they may be of benefit include intraoperative hypotension, aortic stenosis, during cyanotic episodes in Tetralogy of Fallot, and some obstetric situations. In the setting of sepis, phenylephrine has been compared with noradrenaline in which an initial pilot study found a statistically significant reduction in creatinine clearance and increase in arterial lactate after initiating the phenylephrine infusion. However a subsequent randomised controlled comparison of phenylephrine with noradrenaline/norepinephrine did not show differences in cardiopulmonary performance, global oxygen transport, or regional hemodynamics, although there were only 16 patients in each group3.

Phenylephrine is a direct-acting, predominantly α(1) adrenergic receptor agonist used by anesthesiologists and intensivists to treat hypotension. A variety of physiologic studies suggest that α-agonists increase cardiac afterload, reduce venous compliance, and reduce renal bloodflow. The effects on gastrointestinal and cerebral perfusion are controversial. To better understand the effects of phenylephrine in a variety of clinical settings, we screened 463 articles on the basis of PubMed searches of “methoxamine,” a long-acting α agonist, and “phenylephrine” (limited to human, randomized studies published in English), as well as citations found therein. Relevant articles, as well as those discovered in the peer-review process, were incorporated into this review. Phenylephrine has been studied as an antihypotensive drug in patients with severe aortic stenosis, as a treatment for decompensated tetralogy of Fallot and hypoxemia during 1-lung ventilation, as well as for the treatment of septic shock, traumatic brain injury, vasospasm status-postsubarachnoid hemorrhage, and hypotension during cesarean delivery. In specific instances (critical aortic stenosis, tetralogy of Fallot, hypotension during cesarean delivery) in which the regional effects of phenylephrine (e.g., decreased heart rate, favorable alterations in Q(p):Q(s) ratio, improved fetal oxygen supply:demand ratio) outweigh its global effects (e.g., decreased cardiac output), phenylephrine may be a rational pharmacologic choice. In pathophysiologic states in which no regional advantages are gained by using an α(1) agonist, alternative vasopressors should be sought.

These review articles reinforce my own bias against the use of pure pressors in septic shock, although clearly more clinical research is needed. I am inclined to agree with the reviewers’ concluding statement:

…in all clinical settings, phenylephrine reduces cardiac output, and in most clinical settings has been shown to significantly increase LV afterload. Thus, only in instances in which its regional effects are thought to outweigh its global effects should phenylephrine be used for the treatment of hypotension.

1. The physiologic implications of isolated alpha(1) adrenergic stimulation
Anesth Analg. 2011 Aug;113(2):284-96

2. The clinical implications of isolated alpha(1) adrenergic stimulation
Anesth Analg. 2011 Aug;113(2):297-304

3. Phenylephrine versus norepinephrine for initial hemodynamic support of patients with septic shock: a randomized, controlled trial
Crit Care. 2008;12(6):R143
Full Text available here

9 thoughts on “Why I don’t give vasopressors in sepsis”

  1. Great post Cliff!
    You raise a lot of interesting questions about how we judge the adequacy of our resuscitation. I completely agree that we can really be mislead when we rely simply on the numbers instead of judging organ perfusion. I personally believe that MAP, SCVO2, serial lactates and CVP all have a role to play but all have limitations. In some circumstances I think that the patients mentation serves as an excellent indicator of perfusion and progress of resuscitation, and in select cases I defer early intubation so I can watch mentation (though many of these patients will still require intubation as adequate fluid resus plus capillary leak lead to pulmonary edema).

    In terms of the vasoactive drug of choice I think you will get pretty widespread agreement in your view of choice of vasoactive drug. My personal use of phenylepherine is mainly as a push dose pressor used to temporize (and at the expense of peripheral tissue perfusion) to allow critical procedures such as intubation, central line and chatacholamine infusion to be performed and set up.


  2. My friend, excellent stuff as always! A few small quibbles…
    I use the term pressors b/c it is embracing of all of these drugs: there are inopressors like epi and norepi and pure vasopressors, such as phenyl. Vasoactives sort of ignores a important non-vaso site of action: namely the heart. But this point is merely pedantic.

    Of more importance is the one role phenylephrine may have in septic shock, which may be the tachycardic septic patient. They ostensibly have a good deal of circulating inopressors already, and may benefit from the pure alpha. I too was fascinated by the two alpha articles you mention in the anesthesia lit. Neither of them contradict my belief that in a profoundly afterload reduced patient, a pure alpha will certainly improve CorPerfPress ansd therefore CO, so long as they take profound vasodilation, to less vasodilation in the setting of adequate vascular filling. Obviously a misuse of phenylephrine will result in a worse pump than a misuse of the inopressors.

    Last, push-pressors shouldn’t preclude inopressor use. Push-dose epi and if I can find some more literature push-dose norepi will allow you the best of both worlds.

    I’d love to hear your thoughts.


  3. I was hoping Your Excellency would comment on this post Scott!

    My post was deliberately provocative because I want people to (1) question some of the dogma surrounding vasoactive drugs and (2) think about the pathophysiology of whatever shock state is present prior to prescribing them. I agree with all your points, and will briefly reply to them:

    ‘Pressor’ can of course mean anything that raises the BP, and might be an appropriate term (although will often exclude dobutamine). However ‘vasopressor’ (and sometimes pressor) can (and in medical dictionaries, articles, and textbooks more usually does) mean a vasoconstrictor. A typical example is here. I take your point about the limitation of the word vasoactive. Does anyone have a more comprehensive term for ‘drugs that affect circulatory pump and/or vascular physiology’?

    This sounds like a stupidly pedantic obsession with terminology, but I believe it influences the thinking about how the drugs work. I meet many physicians who believe that dopamine and noradrenaline do exactly the same thing, or more worryingly, that phenylephrine/metaraminol are satisfactory substitutes for noradrenaline (even as infusions). I think that is principally because noradrenaline seems still to be described in teaching as a pure alpha-agonist, which it isn’t.

    Coming onto my point (2), your example of supporting coronary perfusion pressure in a vasoplegic patient is a beautiful example of what it would be great to see more of. A physiology-targeted rather than number-targeted intervention based on clinician cerebration. Delicious! I’ve seen too many patients with tension pneumothoraces or exsanguination treated with metaraminol to condone a simplified ‘hypotension = needs pressor’ approach.

    I do buy into the idea of ‘push-pressors’ as a time-buying, life-preserving intervention in the ED or settings where vasoactive infusions are necessarily delayed or difficult. My push-pressor is adrenaline / epinephrine 1 ml from the 1:10 000 minijet added to 9 mls 0.9% saline, giving 0.5-1ml (5-10 mcg) at a time. I haven’t yet seen a need to replace that with a drug more limited in its mechanism and less familiar to ED staff, but I am a ‘keep it simple when the poo’s approaching the fan’ kind of bloke.

    Cheers mate – thanks for contributing to the site


  4. Hi Cliff
    Interesting provocation.
    I am in your camp when it comes to push dose adrenaline at a 1:100000 dilution. Works well and I figure nature has a good reason to give us adrenal glands

    We don’t carry phenylephrine in our retrieval packs and after reading this post I can see why.

    I totally agree with the idea of not simply trying to fix the numbers and giving a drug to try to fix hypotension. Lady Diana’s case where a dopamine drip was started en route to the hospital is one example of that issue. Of course looking for an obvious reversible/treatable cause of hypotension is preferable but as you know in the prehospital setting this can be very challenging. The classic example is the comatose trauma patient with obvious head injury and hypotension. What is bleeding related and what might be neurogenic? How can you tell really without ED base investigations? sure prehospital USS has added another tool in our retrieval box but still it can be very challenging and often you have to treat for both :give volume expanders as well as drugs. The French SMUR system has a protocol for this routinely.
    And in regard to suggesting that we should be inserting more central venous lines in prehospital patients so we can run catecholamine drips more often, I am sure you are aware and have performed peripheral infusions of adrenaline and even noradrenaline for the often short trips back into the base hospital. I consider a short retrieval trip to be under 60min. Dilute noradrenlaine drips are fine for a few hours. peripheral adrenaline drips are totally fine..it seems silly to be giving adrenaline SCI and IMI and we get worried about a peripheral IV infusion of it !
    Dobutamine you can run via peripheral line. I don’t use dopamine prehospitally but if I did it would be the one agent I would try to get a central line or alternatively use an IO for a few hours. ..bit cautious recently as you are probably aware of recent case reports in BMJ and resuscitation journals about compartment syndrome in IO use.
    Generally I try to avoid putting central lines in prehospital setting or even retrieval from smaller hospitals and clinics. I think they are best placed in the receiving ICU.
    Your thoughts, sir?

  5. Welcome back Minh!

    For pre-hospital patients I use peripheral iv or io. I don’t give vasoactive drugs to pre-hospital trauma patients.

    For interhospital missions the vast majority of our patients need good critical care, not time-critical surgery, so if they need a CVC we’ll put one in, using similar equipment and the same sterile precautions they’d get on an ICU.

    Agree peripheral vasoactive infusions are an option, and I gather are used often in North American services, but not (dare I say it?!) standard of care in Australia – we would employ this option as a back up if US-guided CVC failed (which is yet to happen).


  6. Thanks for the provocative post and interesting comments that have followed!

    I share the concerns of the above clinicians. It may come across as a little pedantic worrying about terminology for what to call catecholamine/non-catecholamine sympathomimetics, however, I would rather be a little pedantic then have to watch the unfortunately frequent occurrence of watching people with little experience in critical care having a knee jerk neuronal response of hypotension = start inotropes/pressors. Often with little thought given to the most appropriate one or combinations needed to address a septic patient with often complex pathology where a number of parameters such as filling status, vascular tone, cardiac contractility/septic cardiomyopathy, RV vs LV failure etc need to be addressed.

    Interestingly enough I don’t think any amount of fluids or dopamine would have probably saved Lady Diana. A combination of long pre-hospital time and the presence of a rare and often fatal blunt chest injury ( http://www.sjtrem.com/content/17/1/64 ) led to her ultimate demise.

    I certainly agree with the case made by Cliff for a well sited CVC in an appropriate environment in a non-time critical transfer for infusions. IO are brilliant for most circumstances when needed, although I share some of Minh’s concerns over compartment syndrome ( http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3094012/?tool=pubmed ). Recently I had a case of a 2 year old with severe sepsis/septic shock and difficult PVC/CVC access despite the use of US. They had a correctly sited Cook IO catheter in their tibia and had multiple uncomplicated fluid/blood boluses and dopamine/norepinephrine infusions through it. After a few hours of resuscitation a RIJV CVC was sited and the IO was removed, but unfortunately the child went on to develop compartment syndrome of the same leg and a hyperkalaemia related arrest. However, I suppose rather alive with compartment syndrome then dead with pristine limbs.

    Thanks for all the good work and posts.

    Kindest regards,


  7. HI Cliff
    I am sure you would consider giving noradrenaline infusion to a trauma patient with neurogenic shock, would you not?

    In that setting I would just start a peripheral drip of dilute noradrenaline rather than trying to get a CVC in the prehospital setting. I have had too many cases of inserting a CVC in, then the receiving ICU just rips it out and puts another one in as the first one was not guaranteed to be put under “proper” conditions! You can resus pretty much anything and anyone using a combination of PIV and IO.

  8. Hi Minh

    We don’t carry noradrenaline in our pre-hospital packs. The reality is most neurogenic shock (in our polytrauma, as opposed to isolated cervical spine, patients) is not recognised as such in the field and patients often receive fluid first line (even blood inappropriately). We could give push-dose adrenaline.

    I don’t consider the fact that the receiving ICU team might change the line to be a contraindication to central venous access if I think inserting one will benefit the patient. They’re usually useful for reliable access for multiple infusions as well as the catecholamine infusion, often in patients with crappy peripheral veins.

    Our reason for sending our critical care team is to reduce ‘time to meaningful intervention’ – most of these in non-trauma patients (and actually many in trauma patients) can be provided at the referring site by us. I know you agree with this so I guess our only point of debate is how ‘meaningful’ we consider the CVC to be. That’s a bit patient-dependent but I must admit I certainly put them in less urgently in the ED thanks both to noninvasive sonographic assessment of pre-load responsiveness and the EZ-IO.

  9. This just out: an animal study, but is of interest:

    Background: Myocardial depression is a frequent event during septic shock and may mimic a cardiogenic shock state with decreased cardiac output. Nevertheless, data are scarce regarding the myocardial effects of vasopressors used to treat hypotension. In this study, the authors compared the effects of three commonly used vasopressors acting on different adrenergic receptors on myocardial function in a rodent model of septic shock, as explored with conductance catheter and positron emission tomography.

    Methods: Septic shock was induced in rats by peritonitis. Eighteen hours after septic insult, vasopressors were titrated to increase mean arterial pressure by 20% compared with baseline values.

    Results: We observed that peritonitis was associated with arterial hypotension and systolodiastolic dysfunction. Norepinephrine and epinephrine improved mean arterial pressure, cardiac output, and preload recruitable stroke work, a load-independent measure of systolic function, as well as diastolic function and ventriculoarterial coupling. Heart rate, myocardial oxygen consumption, and arrhythmia incidence were furthermore increased in the epinephrine group. Conversely, phenylephrine, a peripheral [alpha]-agonist, exhibited deleterious effects on systolodiastolic function and ventriculoarterial coupling. Conductance catheter and positron emission tomography yielded identical results with regard to myocardial function evolution under vasopressor treatment.

    Conclusions: Phenylephrine, a drug without [beta]-1 effects, was associated with decreased ventricular performance and ventriculoarterial uncoupling, whereas epinephrine and norepinephrine improved global hemodynamics and myocardial function in severely hypokinetic and hypotensive experimental septic shock. Nevertheless, epinephrine was associated with increased myocardial oxygen consumption. Thus, norepinephrine appears to be a more reliable and safer strategy as a first-line therapy in this particular setting.

    Anesthesiology. 2012 May;116(5):1083-1091

Comments are closed.