Humoral regulation of blood volume and flow

This chapter makes brave attempts to understand Section G5(iv) of the 2017 CICM Primary Syllabus, in which the Court of Examiners expects the exam candidate to "explain the humoral regulation of blood volume and flow". And of course when they say "flow", surely they also mean "pressure" and "resistance". An when they chose the term "humoral", these sagely elders were probably referring to the endocrine control mechanisms which maintain the extracellular fluid volume and vascular tone, rather than the effects of phlegm and yellow bile on the patient's activity and temperament. Restating this syllabus item in a way which facilitates understanding, this chapter will briefly cover the hormonal systems which influence the circulation.

Though some of these systems have been mentioned individually in the college exams (eg. the RAAS, in Question 6 from the second paper of 2011), thankfully nobody has so far been asked to put it all together into a summary statement. There are several systems involved and most people know them all very well by the end of their exam preparation, which makes questions like that very dangerous - without preparation, one may waste their time writing every possible thing they know. Clearly some sort of system is called for. 

For lack of a better system:

The best sources for this sort of thing are actually endocrinology textbooks. For example, an excellent chapter from the 2018 Principles of Endocrinology and Hormone Action by Mannelli et al (Ch.23, p.611-625) would by itself be enough to answer anything they could possibly ask on this topic, and that whole book appears to somehow be available for free in its entirety. In case of some sort of brutal crackdown by Springer-Verlag,  Chopra (2011) will probably remain available through PubMed. For each hormone system here, some individual references are listed in the text, and there is also a short summary available in the Body Fluids section which compiles everything specifically relevant to the control of extracellular fluid volume. 

Sympathetic-adrenal system

As the cardiac reflexes and vasomotor centres have already been discussed elsewhere, the neuro part of this neurohormonal system will only be discussed in passing here. The hormonal element here refers to the release of adrenaline from the adrenal glands. The vasomotor centre (specifically the rostral ventrolateral medulla) sends projections to preganglionic sympathetic neurons in the intermediolateral column of the spinal cord. These neurons, in turn, innervate the chromaffin cells of the renal medulla, so called because they contain dye-absorbant granules full of acidic peptides and catecholamines. The whole medulla can therefore be described as a highly modified sympathetic ganglion, as these chromaffin cells are basically neurons which release their catecholamine neurotransmitters directly into the circulation instead of into the waiting receptors of a neighbouring neuron. In a way, this is still "neurotransmission", except the "synaptic cleft" separating the synapses is the entire extracellular fluid.

Cardiovascular reflexes which trigger the activation of these sympathetic efferents,  the effects of adrenaline and the subtypes of different adrenergic receptors are discussed in greater detail elsewhere, and so it would be pointless to revisit that material here unless something substantially new was being discussed. To simplify revision, this brief summary will suffice:

• Stimulus: hypotension, hypovolemia, hypoglycaemia
• Sensor: various baroreceptors in the carotid sinus, aorta and atria, as well as descending input from the cortex (eg. some emotional trigger for a fight-or-flight response)
• Afferent: glossopharyngeal and vagus nerves, thalami, cerebral cortex
• Efferent: sympathetic nervous system, descending from the rostral ventrolateral medulla via the intermediolateral column of the spinal cord. Specifically, the adrenal glands are innervated by the greater splanchnic nerve, which supplies fibres from T5-T9  to the coeliac plexus. Preganglionic T7-T9 fibres project to the chromaffin cells of the medulla, and these are cholinergic fibres (Parker, 1993)
• Effector: Adrenal chromaffin cells, modified nerve endings which secrete catecholamines (adrenaline and noradrenaline) directly into the circulation
• Effect: vasoconstriction of peripheral circulation (particularly cutaneous),  vasodilation of some muscle vascular beds, redistribution of splanchnic blood flow

In addition to catecholamines, the adrenal medulla secretes all sorts of other bioactive molecules, which are described in glorious detail by Eiden & Jiang (2018). Without going down that particular rabbit hole, the author will instead tease you with the names of these substances (pancreastatin, vasostatin, catestatin, adrenomedullin, and BAM-22P, which stands for "bovine adrenal medulla docosapeptide"). Their endocrine functions are so numerous and diverse that even usually reserved endocrinology authors tend to use the term "protean" multiple times in the same paragraph while discussing these molecules (for one example, BAM-22P is a potent opioid receptor agonist). There is, of course, no possible way for any of this to ever be examinable, and to write anything more about it would be to squander the reader's time.

Like with the rest of the sympathetic nervous system, the adrenal glands maintain a constant baseline level of adrenaline and noradrenaline secretion (in fact usually there's more noradrenaline being produced than adrenaline). With stimulation, for example exercise or stress, this secretion can be ramped up significantly. Kjær et al (1987) determined that the serum levels of circulating catecholamines increase exponentially with work intensity. The authors convinced a group of healthy young subjects to madly pedal on a cycle ergometer while having their subclavian blood tested for serum catecholamines. The original data are exciting enough to be reproduced here in (more or less) their original state:

As one might imagine, the secretory function of each adrenal gland is going to be a pretty unique and individual thing, and so it is unremarkable that the results were very varied between individuals. Also the "healthy volunteers" recruited by Kjær et al were intentionally a rather diverse group, some of whom were trained athletes and others were sedentary randoms (as you can see, one wuss dropped out after literally 2 minutes of pedalling). What is remarkable is that the serum catecholamine levels skyrocketed within minutes and reached extremely high levels, up to 32 nmol/L for noradrenaline.

Of course, ICU patients are usually not trained athletes, nor are they usually exercising maximally on cycle ergometers. More often, they are just calmly dying of shock. What of their catecholamine secretion? Kajihara et al (1977) had to test this in haemorrhaging dogs, who literally bled to death. As the end approached, serum adrenaline levels increased tenfold, and postmortem analysis of chromaffin cells showed only empty vesicles and autophagic vacuoles, a record of the sympathetic nervous system emptying its reserves in desperation.

In humans, the neuroendocrine response to shock is harder to study, as one would have to withhold lifesaving noradrenaline from them in order to measure its endogenous secretion. Benedict et al (1978) had to sidestep this issue by enrolling only shocked patients who did not require vasopressors following their ICU admission, which admittedly selects for relatively "healthy" shock. Unfortunately their method of presenting data was rather messy, including all the patient's data points and using different graphical scales, and so the original data could not be reproduced here, but with some effort a couple of representative patient data sets could be extracted and reproduced here across the same coordinates for comparison:

As you can see, the brief adrenaline peak in haemorrhagic shock was approximately 60 times the normal level (for the record, the normal ranges would be 0.25 and 1.0 ng/ml for adrenaline and noradrenaline). As haemorrhagic shock resolves more rapidly with the replacement of lost blood volume, normality is quickly restored, whereas septic shock simmers over a slow fire. 

With all these serum concentration values being thrown around, one cannot help but wonder - how does this adrenal catecholamine secretion compare to the dose of noradrenaline usually administered to septic patients? Turns out, rather favourably. In a study by Qualha et al (2014), the investigators gave a decent dose of noradrenaline (0.50 mcg/kg/min) to septic children and measured the serum levels for purposes of calculating clearance. Serum noradrenaline increased from a baseline value of 0.54 ng/ml up to 3.75 ng/ml on average, i.e. achieving roughly the same concentration as the septic patients from Benedict et al (1978) were generating on their own. The reader is reminded that the patients from Benedict et al were rather unremarkable haemodynamically, i.e. in order to be enrolled into that study they had to have been admitted to ICU for monitoring but never required noradrenaline.  In other words, in septic shock, your adrenal glands can produce enough noradrenaline to match a drug infusion pump going at 35ml/hr, for a standard 6mg/100ml prescription.

Renin-angiotensin-aldosterone system

There will probably be no better time to discuss this system, as it does not appear in the CICM syllabus anywhere else, and there will probably be no point of mentioning it in the endocrine or fluid sections. It does however appear in the CICM First part exam, as Question 5 from the second paper of 2021, which was failed by 76% of the candidates. Thus, the bulk of this chapter will be dedicated to this system of "humoral regulation of blood volume and flow". Most people will have a sound grasp of it already, as it tends to be well emphasised by medical school curricula. For the same reason, you could find a satisfactory explanation of this subject by grabbing literally any physiology textbook. Or, if for some reason you didn't want to do that, you could look at Fyhrquist & Saijonmaa (2008), or Atlas (2007) who approaches the topic from the viewpoint of pharmacological targets. And though it might seem pointless to add another redundant review where there is already no shortage of peer-reviewed resources, one should be reminded that "pointless and redundant" are the words of our house. 

Anyway. To go through the familiar steps of the RAAS enzymatic signalling cascade:

• Everything starts with renin. Renin is a large-ish 37 kDa enzyme which is synthesised by juxtaglomerular cells of the renal cortex (Persson, 2003). It is produced by cutting chunks out of prorenin, a slightly larger precursor protein (46kDa). Renin remains in storage vesicles of the juxtaglomerular cells, waiting to be released by a range of stimuli (which will be discussed below). Its release and serum concentration level is the rate-limiting step in the pathway of RAAS activation, whereas its substrate is present in the blood at relatively high concentrations, around 1.3mg/L.
• Angiotensinogen is the substrate for renin. Angiotensinogen is a huge wasteful precursor molecule produced by the liver, made up of 485 amino acids and weighing about 62 kDa.  Wasteful, because when renin cleaves it, the product (angiotensin-I) is a relatively tiny peptide, only 10 amino acids long. That leaves behind a huge protein, des(AngI)AGT, as a completely useless byproduct of this reaction, and nobody has any idea of what happens to it or what it does (Lu et al, 2016). 
• Angiotensin-I is an inert decapeptide which also appears to have no physiological activity, apart from apparently stimulating catecholamine release in high concentrations. Its main role is to be the substrate for ACE, and to appear in flowcharts of the RAAS. 
• Angiotensin-converting enzyme (ACE) is the thing blocked by all the whateverpril drugs. As one might imagine for such a high-profile drug target, much has been written about it and its properties are very well described. A good representative reference is this review by Coates (2003). In short, it is ancient (Drosophila have six copies of the homologue gene) and probably involved in many things. Effects of clinical notoriety, apart from making RAAS components, include the degradation of bradykinin, which means that ACE-inhibition can give rise to an undesirable surplus of bradykinin. It is mainly said to be expressed in lung tissue, and on endothelia in general, but enough of it is shed into the bloodstream that concentrated plasma (FFP) can be used to treat ACE-inhibitor angioedema.
• Angiotensin-II is produced from angiotensin-I by ACE.  ACE creates this octapeptide by cleaving a histidine and a leucine from the parent molecule. The result is angiotensin-II, a potent vasoconstrictor. It works by binding to Gq protein-coupled receptors on vascular smooth muscle, facilitating an IP3- mediated increase in intracellular calcium levels, and therefore vasoconstriction. It appears to be about twice as potent as noradrenaline: Rose et al (1960) determined that the same pressor response was seen with 1μg of angiotensin-II and 2μg of noradrenaline. It is a short-lived molecule, being degraded rapidly by endothelial angiotensinases with a halflife of around 30 seconds. Apart from acting as a vasoconstrictor, it has a host of other effects:
  • Stimulating the release of vasopressin
  • Stimulating the release of aldosterone
  • Increased Na⁺/H⁺ exchange in the proximal tubules, thus sodium retention and acid excretion
  • Increased sensation of thirst
  • Increased sensitivity to catecholamines
• Angiotensins are legion. Well, there are a couple more. In addition to angiotensin-II, two other angiotensins (III and IV) are formed by the sequential removal of even more amino acids from the parent molecule, and their roles are unclear. In fact a CICM exam candidate would probably be better off not mentioning them at all. Just forget that they exist.
• Aldosterone release and vasopressin release which occur as the result of angiotensin stimulus also leads to more vasoconstriction as well as more salt and water retention and will be discussed in more detail below. 

So, the main effect of the RAAS is to increase blood volume and blood pressure, and systemic renin release is the main trigger for this cascade. In addition to this, there is also some peripheral capacity to create angiotensin-II, which has some sort of regional paracrine effect. Atlas (2007) called this the "tissue RAS" and remarked that it probably has different effects in different organs, some of which are probably related to cell growth and differentiation (eg. in the heart).

From the above, it stands to reason that renin release should be stimulated by low blood volume and low blood pressure. That is exactly what we see. There are two main mechanisms for renin release:

• Systemic hypotension: the normal baroreceptor responses to low blood pressure stimulate the sympathetic nervous system, which in turn stimulates the juxtaglomerular cells. There are β-1 receptors on the juxtaglomerular cells, which respond to direct adrenergic innervation as well as circulating catecholamines. Then, via increasing intracellular cAMP, protein kinase A mediates the degranulation of renin-containing cells. In fact, anything that ends up increasing cAMP (prostaglandin I2 and E2, milrinone, theophylline) will also produce this effect.
• Renal hypoperfusion: this is a response mediated by some sort of mysterious "renal baroreceptor" mechanisms.  In 1964 Skinner et al were unable to characterise it any better than "a system within the kidney that responds to change in perfusion pressure by altering the rate of renin secretion", and 45 years later Gomez & Lopez (2009) were still asking, "who and where is the renal baroreceptor?" At this stage, it remains unclear. Whatever the mechanism here (connexins? Nitric oxide?), there is clearly a relationship between renal perfusion pressure and renin secretion. Behold, a graph from some dog studies (Finke et al, 1983), which demonstrates that a MAP of about 85 is roughly the threshold for renin release:

  

• Salt depletion can also act as the stimulus for renin release. About this, nobody knows nothing; or as Schweda (2015) puts it, "the understanding of the signaling pathways that transfer changes in salt intake to appropriate adjustments of renin synthesis and release is incomplete." Somehow, by some unknown mechanisms, decreased salt intake produces the release of renin, and vice versa. The macula densa is implicated, as well as several other renal vascular and parenchymal regions. It is probably not essential to unravel that thread for the readers, because ultimately after decades of research it appears to have led nowhere. 

Apart from salt depletion or excess, the magnitude of the renin response is downregulated by several other factors, listed by Kurtz (2012). These are:

• ANP secretion (promotes natriuresis rather than sodium retention)
• endothelin (increases blood pressure)
• angiotensin II (negative feedback mechanism)
• Increased blood flow to the juxtaglomerular cells

In summary, in case one ever needs to represent this in the form of a diagram, the simplest most minimalist form would have to be this:

This whole thing seems to play out over a very short timeframe. For example, haemorrhagic shock leads to an almost immediate reaction from the RAAS. Michailov et al (1987) extracted about 20ml/kg of blood out of anaesthetised dogs and observed that over the 30 minutes of this haemorrhage, the renin and angiotensin-II levels in the blood increased by about four times. That this machinery is powerful in its own right, sans sympathetic help, was demonstrated by Brough et al (1975). Animals rendered areflexic by spinal lignocaine were still able to mount an impressive haemodynamic response following rapid haemorrhage. In the diagram below (slightly modified from the original), haemorrhage was simulated by rapidly withdrawing blood from the carotid artery, over about 30-40 seconds. Over the subsequent 30 minutes, a rather impressive recovery occurred, all without baroreflexes or sympathetic vasoconstriction.

Considering the extremely brief serum half-life of angiotensin-II, it makes sense for this system to be rapidly reactive. However, this brings the question: what use is there for a redundant method of quickly regulating blood pressure? Surely the sympathetic nervous system does exactly the same thing? Why all these extra steps?

Well. One may answer that unlike the sympathetic nervous system (which merely redirects the flow of blood from place to place), the RAAS is able to act in the defence of volume and extracellular sodium (the milieu intérieur, as Fournier et al had put it). The autonomic nervous system has no mechanism by which to increase the retention of fluid and sodium. Moreover, the RAAS appears to be more ancient, in evolutionary terms. It is present in even the lowermost vertebrates; for example, lampreys, who do not have functional sympathetic ganglia, secrete angiotensin-II in response to haemorrhage. There are ACE homologues in flies, and renin-like proteins in leeches (who have an open circulatory system, i.e. their blood is not even in vessels). In short, the RAAS may have existed in some form for much longer than even peripheral resistance vessels, let alone sympathetic baroreflexes. 

Aldosterone

Aldosterone is really all thing to all people, and an endocrinologist would bring up different properties when asked what it does, as compared to a cardiologist, intensivist, or beautician. As the length of this discussion already exceeds the acceptable limits for online discourse, the focus will remain on the role of aldosterone in the defence of blood volume and blood pressure, rather than all the other wonderful things it does. Good reviews by Briet et al (2010) and Briet & Schifrin (2013) were used to fill and organise this section; the first is paywalled but the second is not. 

• Aldosterone is a steroid hormone produced by the adrenal zona glomerulosa, which is the outer area of the adrenal cortex. Bollag (2014) gives a good account of its synthesis and the regulation of its secretion. The main point to remember is that it is not stored in any sort of vesicles, i.e it is secreted by these cells as soon as it is manufactured there. This has two implications: one, that the regulation of aldosterone secretion must occur at the level of its biosynthesis, and two, that an aldosterone response to a stimulus will not be massive and rapid. It is not as if the zona glomerulosa cells will be able to abruptly empty their stores of prefabricated aldosterone, because there are no such stores. The effect would take a while to develop anyway, because rapid is not how steroid work. They slowly make their way to intracellular (nuclear) receptors, where they quietly subvert the machinery of protein synthesis. 
• Triggers for the release of aldosterone are :
  • Angiotensin II release
  • High serum potassium
  • Direct effect of the adrenocorticotropic hormone (ACTH)
• The zona glomerulosa is the "sensor" organ for aldosterone release. Aldosterone is the second "A" in RAAS, and its synthesis is stimulated by angiotensin-2. This is mediated by the A1 angiotensin receptor on the surface of zona glomerulosa cells,  which is a G-protein coupled receptor. ACTH also binds to a G-protein coupled receptor (the melanocortin receptor type 2 or MC2R, in case you are interested). The potassium response is more interesting. Potassium is directly sensed by these cells by means of a fascinating purely electrical phenomenon. Zona glomerulosa cells are highly sensitive to changes in potassium concentration mainly because they have highly conductive membrane potassium channels, and because their T-type voltage-gated calcium channels are rewired with a hair-trigger: their threshold for opening is very close to the resting membrane potential (Spät, 2004). They go off at the slightest provocation. Thus, if the potassium concentration changes even slightly, the membrane potential also changes, the calcium channels open, and calcium enters the cells where it can act as a second messenger molecule to mediate aldosterone synthesis. 
• Effects of aldosterone on blood volume can be summarised as "it increases it".  Aldosterone interacts with a mineralocorticoid receptor in the cells of the distal tubule and increases the expression of a luminal sodium channel which then promotes the reabsorption of sodium and the excretion of potassium. It does something similar in the colon, increasing the reclamation of sodium from the stool. Given how sodium can only really distribute into the extracellular fluid, this move tends to produce an expansion of the extracellular fluid volume. 
• Effects of aldosterone on blood pressure can also be summarised as "it increases it". There are probably two main arms to this effect. One is the indirect effect of increasing the blood volume, which tends to increase the cardiac output, and plays itself out over days.  The other effects are a central vasopressor effect and a peripheral vasopressor effect, neither of which appears to be classically "steroidal", insofar as they do not seem to involve the synthesis of new proteins, and do not follow the normal "genomic" timeframes. Healthy volunteers whose forearms are infused with aldosterone experienced an elevation of regional peripheral resistance over the course of about ten minutes, which could not have been due to any sort of intranuclear-receptor-mediated shenanigans. Wehling et al (1998) who performed that experiment had no idea of which cellular mechanisms might be mediating this. Later authors suggested that several pathways probably play a role, culminating in an IP3-mediated intracellular calcium increase. The central effect is even more mysterious and complex (the reader is directed to Geerling & Lowy, 2009); among other things it involves modifications of behaviour, hunger, thirst, and baroreceptor function.
• The timeframe of most of these effects is days, rather than minutes or hours as with the sympathetic-adrenal system or RAAS. Apart from the abovementioned rapid vasopressor effect, most of what aldosterone does requires some patience. This is illustrated by an excellent dog study by Pan & Young (1982). The investigators acquired eight dogs and infused them continuously with aldosterone (14 µg/kg/day) for sixteen days, recording cardiovascular parameters as they went. Their data is presented below. As you can see, even after one day there was some increase in blood pressure and volume, but the full effect took about two weeks to develop.
• Horrific maladaptive effects are probably outside of the scope of this primary-exam-oriented chapter, but they should probably be mentioned here, if only to explain why we use spironolactone and eplerenone to block those receptors. In short, excess aldosterone mediates all sorts of unpleasant cardiovascular remodelling and inflammatory changes, well detailed in Briet & Schifrin (2013). 

Vasopressin

In the same way as aldosterone is a huge complex topic all on its own, a discussion of vasopressin could take many directions. This one will focus purely on its effects on blood volume and blood pressure. For those in need of a deep dive into vasopressin, Holmes Landry & Granton (2003) produced two excellent papers (Part 1 and Part 2) to satisfy even the biggest vasopressin weirdo.

Vasopressin (antidiuretic hormone) is a nonapeptide secreted by the posterior pituitary. A dated article by Zimmerman & Robinson (1976) is probably still the best explanation of where and how this happens. In short, the synthesis of vasopressin actually occurs in the magnocellular neurons of the hypothalamus, specifically in the supraoptic nucleus and the paraventricular nucleus. That is where the cell bodies of the neurons lie. Vasopressin manufactured here is packaged into granules and then transported by axonal flow into the posterior pituitary, where the granules are opened and the vasopressin is secreted into the systemic circulation via the pituitary portal system, specifically via the short portal veins which connect the anterior and posterior pituitary. There are granules involved and therefore there is a small, finite store of preformed vasopressin available there, but it runs out quickly. 

The stimuli for vasopressin release are hypotension and hyperosmolarity. Hypotension is sensed by arterial baroreceptors in the usual way, and when the nucleus of the solitary tract is responding to this stimulus, one of its efferent projections extends to the hypothalamus to stimulate vasopressin release. The sensing of hyperosmolarity is a little more specialised:

The sensors for hyperosmolarity-initiated vasopressin release are situated in the hypothalamus. Verbalis (2007) gives the best explanation of how they work. In short, two receptor regions mediate this function: the subfornical organ and the organum vasculosum of the lamina terminalis, thankfully abbreviated as OVLT. Vasopressin-secreting magnocellular neurons are themselves osmosensitive, but their responsiveness is somewhat blunted by their being behind the blood-brain barrier; hence the need for these other organs,  which stick out into the bloodstream unprotected to sample the osmolality.  

These sensors are a fascinating variation on the theme of stretch-sensitive baroreceptors, except that they are mechanosensitive at a cellular level. As osmolality increases and decreases, unprotected cells shrink and swell, which produces a change in their membrane tension. This change is detected by transient receptor potential vanilloid (TRPV) cation channel proteins, which are said to be nonselective cation channels with a preference for calcium. The exact signalling pathway from there remains unclear; all that can be safely said is that animals with OVLT lesions tend to become chronically hyperosmolar (i.e. they stop secreting vasopressin).

From the above, it follows that these sensors should be selective, i.e. the thirst and antidiuretic response should be unequal depending on which solute is used to increase the osmolality.  This is in fact what is observed: the sensors seem to prioritise the solutes which are predominantly extracellular: for example, hypernatremic hyperosmolarity is a very potent stimulus, whereas a very raised urea or a massively elevated glucose does not seem to produce the same response. It is almost as if the main point of this endocrine reflex is to preserve the volume of cells rather than the volume of the circulating fluid; solutes which distribute into all compartments are not viewed as much of an osmotic threat. This makes some sense, if you consider that vasopressin-like peptides are probably the most ancient endocrine regulatory system, traced back to the precambrian Archaeometazoa around the time they developed bilateral symmetry and therefore pre-dating bourgeois novelties like the renal nervous and circulatory systems. 

Effects of vasopressin on blood volume are related to the reason it is called "antidiuretic hormone". Vasopressin binding to V2 receptors at the cortical collecting duct stimulates the increased expression of aquaporin proteins on the surface of the collecting duct, concentrating the urine by allowing water to move osmotically out of the duct lumen. The result is water retention. This works nicely with the sodium retention mediated by aldosterone, and if the two systems play well together, serum sodium remains normal. Apart from this renal effect, vasopressin stimulates all sorts of other more sophisticated changes (eg. thirst) which lead to the same thing: increased water intake and decreased water loss.

One might point out that this effect could easily come into conflict with the reflexive defence of blood volume. If the patient is hypovolemic and hypoosomolar,  the hypovolemia stimulus pulls the pituitary in one direction while the low osmolality pulls in the other. In such circumstances, the defence of the volume takes priority. The patient will continue to secrete vasopressin even though the retention of water will perhaps worsen their hypoosmolar state, all in an attempt to maintain a proper circulating volume. This is the basis of the "hypovolemic hyponatremia". Given enough fluid to restore their circulating volume, these people's vasopressin release will suddenly turn off, and they will rapidly self-diurese several buckets of dilute urine as their sodium autocorrects, frightening their endocrinologist with the pons-melting rate of its rise.

Effects of vasopressin on blood pressure

Vasopressin in the circulation binds to V1 receptors, which are widespread through the systemic arterial circulation (mostly sparing the pulmonary arteries). The density of their expression favours the splanchnic arteries, producing a redistribution of blood flow away from the gut which resembles what the sympathetic nervous system and catecholamines do. V1 receptors are G-protein coupled, and binding to them activates phospholipase-β which produces IP 3 and leads to the predictable increase in intracellular calcium via which every vasopressor seems to exert its vasopressor effect. 

This latter effect is often exploited in the ICU, where vasopressin is used as a second agent to decreased noradrenaline requirements, and on occasion (heretically) as a sole vasopressor. It is usually prescribed as a "hormone replacement" rather than a titratable vasopressor, allowed to infuse at a stable rate over a sustained period of time. The rationale behind this way of using it is the theory of "relative vasopressin insufficiency" (Landry et al, 1997), whereby we expect that the patient with a sustained cardiovascular compromise (eg. sepsis) requires so much vasopressin that their posterior pituitary will rapidly run out of granule stores.  

The recommended dose used for this "hormone replacement" is somewhere around 0.04u/min, or 0.04u/kg/hr.  So, if this is supposed to be a replacement of normal function, one might ask the question: how does this infusion compare to the normal systemic secretion of vasopressin? What is a normal vasopressin level, and how is it supposed to scale with the severity of shock? A study of healthy volunteers by Mann et al (2000) found a serum level of around 5 pg/ml when they were euvolaemic, and which increased to about 30 pg/ml when they were dehydrated to an osmolality of around 304 mOsm/L. 

Of course, these were relatively healthy adults, and the investigators did not let them come to harm during the course of the experiment, because this was the year 2000 and somebody would have raised an eyebrow at a study design that intentionally makes people almost die. Fortunately,  Baylis & Robertson were operating in the ethical vacuum of 1979, and were entirely unconstrained by such concerns. They poisoned their volunteers with trimetaphan until their blood pressure dropped by half (we are talking SBP of 60 or so) and recorded some fairly heroic values:

This corresponds to findings by Wang et al (1988) who recorded a maximum value of about 1000 pg/ml from exsanguinating dogs (following 30ml/kg blood loss, or about 40% of the circulating volume). So, when properly stimulated, a nice healthy neuroendocrine system will crank up its secretion of vasopressin by up to a thousand times.  One might expect the vasopressin requirements of critically ill ICU patients to be in this range, but as you will see this did not turn out to be the case. Sharshar et al (2003) measured the vasopressin levels of 44 patients with septic shock (their data is reproduced below), and found a rather anaemic response:

So even in septic shock, early when the granules still have plenty of pre-baked vasopressin, the serum vasopressin level is rather modest, exceeding 15 pg/ml in only one case. 

How much, then, does the routine sepsis-dose exogenous vasopressin infusion contribute? Turns out, quite a lot when compared to normal septic people, but not very much compared to healthy adults, and certainly nothing compared to the doses used by some of the more adventurous investigators. For example, here's a piece of startling ICU pharmacotrivia. Ebert et al (1986) infused healthy males with 15 and 40ng/kg/min of vasopressin, increasing their serum concentration of vasopressin from 5 pg/ml to 18 and 36 pg/ml respectively, i.e. well in excess of what is usually expected under "natural" circumstances in shock.  But wait: working from the assumption that 1mg of vasopressin contains 530 units, that dose they used amounts to about 30-90 units per hour. This exceeds all normal infusion standards in modern clinical practice, and even the most liberal guideline documents draw the line at 0.6units/kg/hr (which is about 42 units/hr for a 70kg person).

A more normal dose rate would be about 2-3 units power hour.  During the VASST study the investigators measured vasopressin levels of septic shock patients receiving an infusion of about 0.03U/min (1.8 U/hr) and found vasopressin levels increased from a low baseline of around 3 pmol/L up to about 70-100 pmol/L (Gordon & Russell, 2010). At a baseline, in a healthy euvolaemic person, the pituitary maintains a level somewhere around 3-4 pmol/L, which means your own internal rate of vasopressin infusion is about 0.07-0.10 units per hour. 

Fortunately for the victims of Ebert et al, the cardiovascular effects of vasopressin appear to be completely unrelated to its dose. And its maximal effect is only apparent in patients with shock. Those healthy young men in the 1986 experiment were completedly unharmed by their 90u/hr vasopressin binge, by the way. Their MAP increased from about 86 to about 94. Those receiving 30u/hr barely had experienced any change in their blood pressure. Dünser et al (2004), in unhealthy completely-falling-apart patients with what the authors described as "advanced vasodilatory shock" also found no relationship between the serum vasopressin level and haemodynamic effect. But at least it had a haemodynamic effect-  vasopressin at "modest" doses up to 4u/hr predictably and significantly decreased the noradrenaline requirements in these patients.  Clearly, something unique and wonderful happens in severe vasoplegic shock which potentiates the effects of vasopressin.

What is that unique and wonderful thing, that restores vessel wall tone in sepsis, but spares a normal circulation? Well, there are probably at least four such things. Landry & Oliver (2001) list references in support of the following hypotheses:

• In shock, there is a relative vasopressin deficiency, and the receptors are all unoccupied (whereas all the other vasopressors are saturating their receptors)
•  Vasopressin potentiates the effects of noradrenaline, which is present in vast concentrations in a shocked patient, but only in minimal concentrations in a healthy euvolaemic patient.
• Vasopressin directly inactivates K_ATP channels in vascular smooth muscle, which is extensively activated in sepsis and which contributes significantly to the vasoplegia (Buckley et al, 2006), but which is fairly quiescent in health.
• Vasopressin decreases the synthesis of inducible nitric oxide synthase that is stimulated by bacterial lipopolysaccharide and inflammatory cytokines, which leads to the recovery of vessel tone. 

So, in short, rather than acting as a vasoconstrictor which brutally imposes its will upon the vascular smooth muscle in a linear dose-dependent manner, in sepsis vasopressin acts mainly to restore vessel tone and catecholamine responsiveness, basically returning vascular reactivity to a point where other vasoactive drugs can be useful in nontoxic doses.

Natriuretic peptides

This topic is well covered in the literature, as natriuretic peptides are turning into commercially attractive diagnostic biomarkers and drug targets. Good representation which pre-dates the corporate taint of Big Peptide can be found in Potter et al (2009) or Baxter (2004). In summary, there are three of them. ANP and BNP (atrial and brain natriuretic peptide) are the sexy interesting hormones, and CNP (C-type natriuretic peptide) is somewhat forgotten as it appears to support long bone growth and other "unappreciated functions". Nobody talks to CNP at parties. 

ANP and BNP are relatively large (28- and 32-amino-acid) peptide molecules which are structurally related (they both have the same 17-aa ring). They seem to be well conserved from the evolutionary perspective, our ANP differing from the rat by just one amino acid (we preferred methionine to their isoleucine). Synthesis of these molecules takes place in atrial muscle. BNP is said to come from the ventricle and ANP is said to come from the atrium, but in fact it appears that both substances can be found in atrial granules, and both tend to be released in response to atrial stretch. In contrast, only BNP is released in response to ventricular stretch and is therefore a more sensitive biomarker for ventricular volume overload. ANP is preformed and stored in granules, whereas BNP is synthesised as needed. Of the two hormones, BNP has a longer circulating half-life (20 minutes, vs 2 minutes for ANP). 

Stimulus for ANP and BNP secretion is basically myocardial stretch, which occurs in response to an increase in preload. There seems to be some baseline level, which increases with any state which puts pressure on the chamber walls (for example, the failure of an upstream chamber). The precise mechanoreceptor process involved in this remains to be explained. Apart from mechanical stretch, Fu et al (2018) list a whole bunch of other (totally predictable) release modulators, such as exercise, hypoxia, cold, angiotensin, endothelin, vasopressin catecholamines and corticosteroids. Natriuretic peptides are released into myocardial capillary blood and end up being washed into the bloodstream out of the coronary sinus.

Natriuretic peptide receptors are located mainly in the renal tubule, vascular smooth muscle, adrenal glands, the heart and the brain.  These are membrane-spanning receptors attached to an intracellular guanylyl cyclase which produces cGMP as a secondary messenger. Downstream effects are then mediated by protein kinase G and phosphodiesterase.

The effect of stimulating these receptors is basically natriuresis and diuresis. ANP inhibits sodium channel function and Na/K ATPase activity in the cortical collecting duct, decreasing sodium reabsorption. It also directly inhibits renin release, interferes with vasopressin binding to its V2 receptors, and directly vasodilates the afferent arterioles of the kidney (Wong et al, 2017). In short, everything happens to increase the excretion of sodium and water, which reduces the cardiac chamber distension. 

Thyroid hormone

It will surprise nobody to hear that the thyroid gland interferes in the control of blood pressure, as it interferes in virtually everything. Hypothyroidism is associated with hypotension and low cardiac output, whereas hyperthyroidism gives rise to a hyperdynamic circulation with an occasional hypertensive crisis. The cardiovascular effects of thyroid hormone are well reported in Danzi & Klein (2003), and it would probably benefit nobody to get into them in any great detail here, considering especially that many of the effects are not particularly well explained and therefore do not lend themselves to a good SAQ or viva answer. In summary, it has direct and indirect effects:

• Peripheral vascular vasodilation, which results in a reflex increase in cardiac output
• Increased vascular reactivity to catecholamines
• Increased vascular reactivity to angiotensin-II
• Increased blood volume, probably for multiple reasons (eg. because RAAS)
• Increased synthesis of renin and angiotensinogen

Cortisol

Cortisol, the major human glucocorticoid, is produced by the adrenocortical zona fasciculata. The total list of their functions is massive and complex, including inter alia some influence on the maintenance of blood pressure and blood volume. Corticosteroid release is triggered by various forms of stress, a spectrum which incorporates shock and hypovolaemia. They then produce a restoration of blood pressure and blood volume by a variety of mechanisms, which are detailed in Whitworth (2005).

If no shock or hypovolemia is present, corticosteroids will produce hypertension and volume overload. When Whitworth et al (1984) famously filled healthy volunteers with cortisol, they found that after five days their blood pressure increased by about 20mmHg on average, and their bodies gained about a litre of fluid. The latter effect is owed to their cross-reactivity with mineralocorticoid receptors, which increases the renal reabsorption of sodium and decreases the renal excretion of water. The vasopressor effect is probably multimodal and indirect, i.e. mediated by an increase of catecholamine receptor sensitivity and interference with nitric oxide-mediated vasodilation.