This chapter is relevant to Section G8(i) of the 2017 CICM Primary Syllabus, which asks the exam candidate to "understand the detailed pharmacology of inotropes and vasopressors". Noradrenaline is an endogenous catecholamine, a sympathomimetic drug with a strong alpha-1 receptor selectivity. This chapter is a tribute to it, as it is the true workhorse of intensive care, and a drug with which one ought to become intimately familiar. It is the gateway drug to understanding the effects of catecholamines. Surprisingly, past CICM exam papers have only ever asked about it in a couple of SAQs:
Class Vasopressor Chemistry Catecholamine Routes of administration IV Absorption Basically zero oral availabilty due to destruction by brush border enzymes in the gut (COMT and MAO) Solubility pKa = 8.85; water-soluble Distribution VOD = 0.12 L/kg, i.e. essentially confined to the circulating volume; 25% protein-bound Target receptor Noradrenaline is highly selective for the alpha-1 receptor Metabolism Metabolised rapidly and completely by COMT and MAO Elimination Metabolites are renally excreted. Half-life is ~2 minutes Time course of action Very short acting, very rapid onset of effect Mechanism of action By binding to the alpha-1 receptor, noradrenaline increases the release of a secondary messenger (inositol triphosphate, IP3) which results in the release of calcium into the cytosol, and thus enhanced smooth muscle contractility. Clinical effects Increased peripheral resistance, increased afterload, increased blood pressure; redistribution of blood flow from splanchnic circulation and skeletal muscle. Single best reference for further information TGA PI document
Noradrenaline is the prototypical alpha-agonist. The relationship of its structure to its function is discussed elsewhere. The hydroxyl group on its beta-carbon increases its overall potency, and the absence of alkyl substitutes on the amine group gives it a degree of alpha-1 selectivity, but takes away beta-activity.
The Australian variety (“Levophed”) arrives in your hands as a 2ml ampoule, containing 2mg of noradrenaline. It has a pH of about 3.5-4.0. In order to make it isotonic, Hospira thoughtfully provides 8mg of sodium chloride per ml, which equates to a concentration of about 136 mmol/L. Additionally, sodium metabisulfite (0.2mg/ml) is added as an antioxidant; as far as I can tell it plays no real role in the ensuing discussion.
The name "Levophed" and the conspicuous presence of a chiral carbon in the middle (the one with the OH group) should suggest to the reader that this catecholamine has at least one enantiomer. Resisting the urge to launch on a long and pointless digression, it may suffice to remark that most neurotransmitters have chiral structures, and that most of them are maximally biologically potent as the L-enantiomer, with the D-enantiomers occurring rarely in nature and having a much weaker pharmacodynamic effect, or occasionally even toxicity (as with D-Dopa). Thus, under most circumstances, where it is cheap and convenient to do so, pharmaceutical preparations will present mainly the active enantimer (the "eutomer") to reduce the necessary dose by excluding the "isomeric ballast".
So, noradrenaline also has a D-isomer, which is ten times less potent (Lv et al, 2020), not found in the living organism in clinically significant concentrations, and generally not expected to be present in the ampoule. However L-noradrenaline does undergo spontaneous racemisation while in storage, which means it may theoretically lose potency over time, and so the patient may occasionally get some D-enantiomer infused as a part of their routine care. Apart from being much less recogniseable by the receptors, the pharmacokinetics of the D-enantiomer also demonstrate some stereoselective weirdness, including slightly slower elimination, slightly different protein binding and slightly different volume of distribution. For the remainder of this chapter, it will be ignored along sodium metabisulfite, as an inert impurity we are aware of out of the corner of our eye, but not interesting enough to merit any further attention.
Why does it have to be so acidic? Because noradrenaline (and adrenaline for that matter) are degraded by an alkaline environment, particularly at above room temperature. This is one of the reasons the ICU nurses never run sodium bicarbonate into the same CVC lumen as the noradrenaline.
Why do we always seem to prepare it in a bag of 100ml dextrose? It doesn’t matter whether you put it in dextrose or saline, right? Certainly, it seems to be equally stable in either solution. However, the prescribing guidelines for the Levophed product recommend against using pure saline as a diluent, citing as their reason the potential for oxidative degradation in a pure saline solution. Though I cannot find convincing articles to support this recommendation, it certainly seems to be followed around New South Wales. There may be some historical inertia at play here; the first papers (from 1952!) describing the use of noradrenaline used dextrose because the authors were using a dilute concentration of the drug in cardiogenic shock patients, and they were concerned about excess sodium.
One might ask, what is it with adrenaline and noradrenaline, being prepared as 6mg per 100ml? Surely it doesn’t matter? Surely, one can put 100mg per 100ml, or any damn amount? Interesting thought.
The 6 in 100 rule is for convenience of dose calculation. Conventions of inotrope and vasopressor dosing had once dictated strict parameters, in terms of micrograms per kilogram per minute. And indeed, some units still draw up their catecholamine infusions according to the patients weight, so that 1ml per hour equates to 1mcg/kg/min. Similarly, having 6 mg in 100ml equates the rate of 1ml/hr to 1mcg/minute. However, since we really just titrate the rate to its effect, this convention is an anachronism. Dilute your norad any which way you please.
The family of noradrenaline is the catecholamines- specifically, the endogenous catecholamines. Noradrenaline is the daughter drug of dopamine and the parent drug of adrenaline. As catecholamines, they all share certain sympathomimetic properties; they are all poorly lipid-soluble, short-lived, and potent in small concentrations.
It is an intravenous medication. No fool would ever drink noradrenaline. For one, it is rapidly destroyed by brush border enzymes, and none will ever make it into your bloodstream. In that way, it is analogous to the phenylephrine in over-the-counter decongestants.
Even if one did drink some noradrenaline, I expect one would be disappointed by the taste. It comes as a tartrate salt, suspended in a sodium chloride vehicle. These two things alone would likely render it a bitterly sour flavour, resembling aspirin crushed into a shot of gin. Extensive exploration of the literature has not yielded any reports of accidental ingestion. However, this product safety data sheet recommends the foolish ingestor induce vomiting immediately (which makes no sense to me, seeing as one’s stomach is perhaps the safest place for noradrenaline to remain).
One learns early in one’s ICU career that peripherally administered sympathomimetics ought to be limited to weak alpha-1 agonists such as metaraminol, or to arrest situations. Cautionary tales abound. Gross ischaemic bullae bulging from the pages of medical journals warn us not to give this drug through the peripheral circulation.
This was not always the case. In the beginning, noradrenaline was a peripherally administered drug. The Annals of Surgery grant us a curious glimpse at this era, during which peripheral infusions of noradrenaline seem to have been the standard of practice, and the medical journals were awash with correspondence about how best to treat the seemingly inevitable “blanching”- pale patches of ischaemic skin which would appear on these patients’ arms.
The tendency of noradrenaline to extravasate has been well described even in the 1950s; it is not merely a matter of cannula dislodgement. Rather, it appears to be the direct effect of concentrated sympathomimetic on the vasa venorum, the blood vessels in the walls of the vein. Predictably, they contract on contact with noradrenaline, rendering the vessel wall ischaemic – and this can be seen as a blanching of the vein, which gradually extends over the length of the vein. Thus crippled, the vein can no longer maintain the integrity of its wall, and noradrenaline escapes into the subcutaneous tissues, causing necrosis.
With these potential complications, one must mention the fact that peripheral use of vasoactive drugs is undergoing something of a resurrection. PulmCCM have dissected this practice with some attention to detail. The authors of a recent study (Cardenas-Garcia et al, 2015) used peripheral noradrenaline in 730 patients with reasonable safety (no arms fell off). The safety of their technique relied on ultrasound-guided placement of large-bore lines into large peripheral vessels (>4mm in diameter), followed by vigilant monitoring for complications of a relatively short period of infusion. There is therefore a sound argument to be made in favour of short-term peripheral noradrenaline as a rescue therapy in resource-poor environments where no central access is possible, or where central vein cannulators are disastrously clumsy and error-prone. Of course, these environments usually also have access to metaraminol, which can be given peripherally without the risk of tissue injury and which is still an effective vasopressor over a short term.
Noradrenaline has a very short serum half life, measured in minutes. The metabolism of catecholamines is discussed in detail elsewhere; to briefly summarise, two enzymes are important: monoamine oxidase (MAO) and catechol-O-methyltransferase (COMT). Of these, COMT is the most important to the metabolism of intravenous catecholamines.
The ultimate fate of catecholamines is complicated, but the important thing to remember is that the first step of metabolism produces inactive metabolites, which are ultimately ground down into something easily excreted.
Noradrenaline is frequently said to be a “pure” alpha-1 agonist. Some alpha-2 effect is also observed, and tends to mediate its effects on the tiny pre-capillary arterioles. One may examine the effects of alpha-1 adrenoceptor stimulation and the intracellular signalling pathway involved. Suffice to say, the pathway involves the activation of a secondary messenger (inositol triphosphate, IP3) which results in the release of calcium into the cytosol, and thus enhanced smooth muscle contractility.
But this, though interesting, is far removed from the clinical arena. Let us now discuss the effects of noradrenaline in the human organism.
Much of what we know about the haemodynamic effects of systemic catecholamine infusions has come to us from the late forties and early fifties, as a result of very interesting work by the early pioneers of the field.
The most commonly displayed graphical representation of the comparative effects of adrenaline, noradrenaline and dopamine can be found in the famous article by Allwood et al (1963). Allwood and friends infused healthy volunteers with 10mcg/min of noradrenaline (or, about 0.14mcg/kg/min). The resulting data is adapted here without any permission whatsoever.
In these healthy volunteers, noradrenaline increased both the systolic and diastolic blood pressure; the heart rate dropped because of baroreceptor-related compensation. However, these are merely the effects of noradrenaline on healthy volunteers, at modest stable dose rates. As the dose escalates, so the effect of receptor selectivity diminishes.
In 1986, Pang and Tabrizchi had published a lovely paper comparing the dose-response curves of various vasopressors. I will modify their graph for noradrenaline slightly, because instead of mcg/kg/min, on their X axis they used a logarithmic scale and measured (very scientifically, but confusingly) their noradrenaline dose in mol/kg/min.
The dose range for noradrenaline they examined was from 3.0 x 10-10 to 8.0 x 10-9 mol/kg/min. Armed with Wikipedia and basic arithmetic, one is able to calculate that in our old familiar terms, this range is 0.05mcg/kg/min to 1.35mcg/kg/min.
Yes, it's not a very exciting curve. The alpha-1 effects of noradrenaline have a very linear and predictable relationship with the dose, particularly at lower doses. This makes it a nice well-behaved drug to work with
The above graph is extrapolated from animal studies, human data, and (limited) personal experience.
Early dog studies have demonstrated that at relatively low doses (up to 0.5mcg/kg/min) the dose response of noradrenaline is essentially linear. That is to say, a predictable increase in blood pressure will result from a given dose increase. Both the systolic and diastolic pressures rise in unison. The pulse pressure remains essentially the same.
As the dose continues to increase the pulse pressure widens, as noradrenaline begins to act on beta-2 receptors. Its effect on these was cleverly demonstrated by Brick Hutchinson and Roddie, who showed that the same dose of noradrenaline has a more potent vasoconstrictor effect in a beta-blocked human forearm. As these effects manifest, the diastolic pressure decreases.
In short, at high doses, noradrenaline begins to act like adrenaline.
There are limits to everything, and patently the systolic pressure cannot continue to rise indefinitely. At a point, the left ventricle will struggle to overcome the systemic vascular resistance, and cardiac output will decrease, thereby decreasing the blood pressure.
In short, initially there is none. CVP does not vary dramatically in the low dose range of noradrenaline infusion, and over moderate doses it varies inversely with the cardiac output, suggesting that the drop in cardiac output (a loss of forward flow) results in decreased removal of volume from the central venous circulation. However, noradrenaline does affect the venomotor tone in the central venous circulation, and decreases the central venous capacity. The effect is rather vigorous- the authors mention that “two drops” of noradrenaline on an exposed venous surface produced an immediate and vigorous effect on venous smooth muscle.
This has implications for fluid responsiveness. It would make sense that a lax and flaccid central venous circulation will act as a massive capacitance reservoir, absorbing large quantities of fluids before an increase in CVP is measured. A tense and rigid noradrenaline-soaked central venous circulation will be more-fluid responsive, and the CVP will rise in response to smaller fluid boluses. However, I cannot find any experimental evidence in the literature to support this assertion.
Gaal, Kattus Kolin and Ross injected some dog coronary arteries with noradrenaline, and found that initially, it dilated them. This seems to be a transient phenomenon, which these investigators were able to ablate completely with nonselective beta-blockade. In sustained infusion, coronary vascular resistance increases, which seems to be the old familiar alpha-1 adrenoceptor effect.
One may expect little effect from noradrenaline on the cerebral circulation, given that it does not cross the blood-brain barrier. Early studies are available, which investigated living human patients by injecting noradrenaline directly into the carotid artery. This sounds like a nightmarish complication of CVC malposition, and I for one am glad that somebody has performed this experiment intentionally, so when asked “what would happen if I gave norad directly into the brain” I would be able to respond coherently.
In short, even in absurdly high doses carotid pressure only increased slightly; however the intracranial vascular resistance did increase. Vasoconstriction of the cerebral circulation had taken place, and it is thought to be the direct effect of noradrenaline on the small diameter cerebral vessels. King et al had demonstrated that this increase in cerebrovascular resistance results in decreased cerebral blood flow.
The effect of catecholamines on the cerebral circulation is more pronounced under anaesthesia with propofol or volatile agents. Myburgh et al investigated this effect in sheep, and concluded that blood-brain barrier permeability increases must allow catecholamines to influence cerebral blood flow more directly, and hypothesized that injured brains may therefore be more susceptible.
How does this affect regional blood flow? Is cerebral oxygenation affected? In a word, yes, and particularly under anaesthesia. At doses above 0.1mcg/kg/min (that’s about 7ml/hr for a 70kg person) cerebral oxygenation will begin to suffer. One can imagine the implications for the subarachnoid haemorrhage patient in whom one is trying to avoid cerebral vasospasm.
A reflex bradycardia results from the infusion of low dose noradrenaline, as the vagal baroreceptor reflex forces a compensatory slowing of the sinus node. However, as the dose increases, the receptor selectivity decreases. An experiment discussed later in which dogs were overdosed on noradrenaline had found that the reflex bradycardia had become reversed at doses beyond 1mcg/kg/hr. Thereafter, beta-1 mediated tachycardia ensues. In fact so much is this a problem, that the abovementioned Pang and Tabrizchi rat paper had the noradrenaline rats pre-treated with propanolol, so as to ablate the beta-1 effects of noradrenaline.
Is this bradycardia ever going to happen in the sick ICU patient? I doubt if we ever see a picture as pure as these rat studies. Our patients are semiconscious, in pain, losing blood and fluid (or gaining frightening volumes thereof) and marinading in a cocktail of bacterial endotoxin, as well as drugs which interfere with their autonomic nervous system. Whatever catecholamines we infuse will have their effect modified by their endogenous catecholamines, their vagal tone, their sympathetic tone, and all the other complicated stuff.
In short, user experience will vary.
Overall, the cardiac output remains more or less stable throughout the low dose range, and only begins to drop towards the high range of doses. The combination of increased afterload and myocardial damage results in a decrease in cardiac output, even though in the disembodied heart, noradrenaline at absurdly high doses increases cardiac contractility, much as adrenaline does. At low doses, with only alpha-1 effects observed in healthy volunteers the modest cardiac output reduction has been put down to vagal reflexes and bradycardia.
At least in septic patients, RV volumes and RV ejection fraction remained unchanged over the studied dose range.
The effects of noradrenaline on this part of the circulation has been studied in the dog, albeit with boluses of noradrenaline rather in constant infusion. At least the boluses increase the pulmonary arterial pressure. Indeed, the pulmonary circulation is well supplied with alpha-1 receptors, though the smooth muscle there is perhaps somewhat more flimsy than in the systemic circulation. Even so, in the anaesthetised dogs a 2mcg/kg bolus managed to more than double the PA pressure (from about 25mmHg to over 50mmHg).
It is generally said that acidosis decreases the sensitivity of the circulatory system to catecholamines. This appears to be more true for alpha-2 rather than alpha-1 receptors. The tone of small precapillary arterioles is controlled predominantly by alpha-2 effects, and modest acidosis has been shown to reduce the noradrenaline-related change in their diameter, suggesting that in acidotic patients at least this part of the circulation will lose its noradrenaline sensitivity. Oddly, large alpha-1 sensitive microvessels did not seem as affected by acidosis. Still, rat tail artery experiments have demonstrated that in conditions of severe metabolic acidosis, even alpha-1 receptor responses are blunted (whereas vasopressin receptor responses are preserved).
However, there is more to this story. It seems acidosis tends to uncouple the vascular smooth muscle from sympathetic control, which means that in a shocked patient one would start to lose all that useful sympathetic tone. This would translate into increasing noradrenaline requirements, which can be interpreted as noradrenaline resistance. This has been demonstrated experimentally in an animal model of lactic acidosis, as well as in animal models of bacterial endotoxin shock (where E.coli endotoxin produced an acidosis with a base deficit of -20, and noradrenaline responsiveness was greatly diminished).
So, we have been using noradrenaline for shock for over six decades now. And, from reading the early works, it would seem that initially its application was limited to patients with cardiogenic shock (these are all those poor bastards in the 1950s who got it infused into them via peripheral veins). This is still done, and seems to be well tolerated even when you aim for a MAP around 85 (to improve diastolic pressure and coronary arterial filling), without compromise to cardiac output in spite of the increased afterload.
Even though we know that noradrenaline produces coronary vasoconstriction, particularly in beta-blocked individuals, the AHA still recommends it as the pharmacological agent of choice (though not by name) in cardiogenic shock with poor fluid response (level C recommendation).
In short, noradrenaline is a useful tool for increasing arterial blood pressure in cardiogenic shock.
As a staple of ICU, noradrenaline has been employed to boost the saggy blood pressure of post-CABG patients ever since CABGs began. In this population, it performs as expected. It can increase cardiac output by "venous recruitment" (pushing blood out of the central venous system by constrciting those veins). Or , it can decrease cardiac output by increasing afterload. At least in one trial, it seems to have come off second best when compared to vasopressin.
Noradrenaline is a fine choice for septic shock. In the past, dopamine was thought to be the drug of choice for vasopressor support in sepsis, because the vasoconstriction associated with noradrenaline was thought to be deleterious in some way- organ hypoperfusion and whatnot. However, a prospective cohort study of septic patients demonstrated that in fact the use of noradrenaline was strongly associated with a favourable outcome. A later trial confirmed that in comparison to dopamine, noradrenaline is associated with decreased mortality in septic patients (by virtue of fewer arrhythmic events). And the VASST investigators confirmed that it is no worse than vasopressin in the setting of septic shock.
As already mentioned, noradrenaline at doses above 0.1mcg/kg/min will compromise cerebral oxygenation. Which in the setting of subarachnoid haemorrhage is a weird conundrum. Do you protect cerebral regional blood flow by sparing your patient from vasopressors, or do you protect that blood flow by cranking up the vasopressors to squeeze more blood though spasming arteries?
In summary, nobody knows what to do. Neurosurgical literature seems to favour induced hypertension more than intensive care journals. Surveys of neurocritical specialists demonstrate that there is a lack of consensus about management strategies and endpoint goals. When stuck, noradrenaline still seems to be the agent of choice, because ultimately everybody agrees that hypotension is totally counterproductive in these patients.
So you need to keep the blood pressure up, and you don’t know why its low, and you have all these wonderful vasoactive drugs to throw around. Which do you chose?
Well, noradrenaline seems to a be a good solid performer. Let us review the evidence:
A nice 1679-patient randomised multicenter trial had demonstrated that in shock of all causes, noradrenaline has fewer adverse events (and improved 28-day mortality) than dopamine. The subgroup which benefited most (or rather, the subgroup most harmed by the dopamine) seemed to be the cardiogenic shock patients.
Relative contraindications for noradrenaline infusion are listed comprehensively in the MIMS entry for Levophed.
I will briefly reiterate:
Noradrenaline interacts with monoamine oxidase inhibitors, which (obviously) interfere with its metabolism. The duration of its action will be increased. Additionally, noradrenaline may interact with some of the older tricyclic antidepressants (which inhibit MAO to some extent).
We have all seen the scenarios where patients end up on quadruple-strength noradrenaline. The question one frequently finds himself asking is whether this insane dose escalation might somehow be counterproductive.
In 1966, Moss Vittands and Schenk published a series of articles in Circulation Research, of which the first dealt with the haemodynamic effects of noradrenaline. These were studies in anaesthetized mongrel dogs, receiving 1mcg/kg/min, 2 mcg/kg/min and 4 mcg/kg/min of noradrenaline, which incidentally are doses wholly incompatible with normal human survival.
Let us dwell on that for a moment, extrapolating the result to human beings.
Let us take a 70kg person.
Consider that even 1mcg/kg/min equates to 70ml/hr of single-strength noradrenaline in a normal person. In ICU, one most frequently encounters noradrenaline in the dose range between 5 and 20ml/hr, equating to somewhere between 0.1mcg/kg/min and 0.3mcg/kg/min.
Unsurprisingly, as the doses of noradrenaline were escalated the animals developed left ventricular failure and died. At a rate of 4mcg/kg/hr, this sort of left ventricular failure developed after the second hour of infusion.
It is known that catecholamine infusions in excessive quantity cause diffuse myocardial damage. The effects of truly absurd catecholamine doses were demonstrated experimentally by authors who (in 1960) commented on the “mounting frequency of myocarditis due to pressor amine therapy in human beings”. The experimental dogs were infused with noradrenaline at doses ranging up to 10-15mcg/kg/min.
In the abovementioned normal 70kg human being, a dose rate of 10mcg/kg/min would yield 700mcg/minute, … yes that is correct, 700ml/hr by the standard dilution, or 175ml/hr if drawn up to quad-strength.
Think about that. No intensivist, no matter how crazed, would ever infuse a human being with that sort of dose.
However, let us consider what might happen if they did.
The dog study mentioned above reports that in animals receiving 10mcg/kg/min, death occurred within 30-180 minutes, usually of cardiac arrest, cerebral haemorrhage or pulmonary oedema. At half that dose (equivalent to about 85 ml/hr of quad strength noradrenaline in the human) the dogs lasted maybe six hours. The second paper from the series delved deeper into the histological changes of the myocardium following such a noradrenaline overdose.
So, if nothing else, these animal experiments give one an impression of how high one shouldn't go with one’s infusion rate.
Let's say you accidentally overdosed your patient with a bolus of noradrenaline. A sensible reaction to this would be to administer a competitive antagonist of alpha-1 receptors, such as phentolamine. This has been used to treat the infarcted digits of patients who accidentally used the wrong end of an Epi-pen autoinjector (though that was raw adrenaline, one can see the parallels).