Mechanisms which maintain extracellular fluid tonicity

This chapter is a lamentable departure from the central theme of Section I1(i) in the 2017 CICM Primary Syllabus, which is supposed to "explain the distribution and movement of body fluids". It is also vaguely associated with SectionG5(iv), "explain the humoral regulation of blood volume and flow", insofar as the main agents involved are certainly "humoral". There have been no specific CICM First Part Exam questions on this subject, other than those related more broadly to the regulation of body water, and so it was really quite pointless (and possibly harmful) to write this piece into a section titled "Required Reading", as the reader may come to the incorrect conclusion that this is valuable or necessary, when all they really need is a barebones summary like the one below:

Still, they say the writer's role is to menace the public's conscience. In that spirit of unhelpfulness, the reader can be directed away from gainful exam-oriented study towards works like Juul (2012) or Donaldson & Young (2012).  To sabotage them further, they can also be offered From Fish to Philosopher by Homer W. Smith (1961). People trying to keep away from deep rabbitholes will probably be better served by review papers like Bayliss (1987) who summarised the regulatory role of vasopressin with a fair compromise of detail and brevity, or Celia Sladek's highly detailed vasopressin chapter for Comprehensive Physiology (2010) if they are unwilling to maker that sort of compromise.

Extracellular fluid tonicity

The normal homeostatic setpoint for extracellular fluid osmolality of the human organism is generally said to be about 280 mOsm/kg, with an acceptable laboratory range of about 275-295 mOsm/kg. The reader may be interested to learn that this is not a law of nature. There is nothing scientifically "ideal" about this value, and it only applies to terrestrial vertebrates; other organisms do not obey the same rules and tend to find different ranges of osmolality acceptable, whereas yet others could not possibly care less about their extracellular osmolality, and do not feel the need to do anything even remotely osmoregulatory. Many marine invertebrates,  for example, are just happy to be isoosmolar with the sea water, and are referred to as osmoconformers.  They can get away with it - the environment they live in tends to be fairly static - but the fact remains that they have no real mechanism to control extracellular tonicity, and this is a major factor limiting their adaptability to environmental change.

Whereas in the halcyon days of the osmotically stable Cambrian our earliest aquatic ancestors could  frivolously squander their primitive nonapeptide hormones on wanton egg-laying and ejaculation, these fast-paced modern days we terrestrial organisms need them to keep an osmotic homeostasis.  For example, for teleost fishes the range of possible waters spans from basically 0 mOsm/kg (i.e some kind of crystal clear mountain spring) to about 3200 mOsm/kg, which is the briny environment of drying saltwater rock pools in Israel. Some, for example Aphanius dispar the Arabian toothcarp, can actually survive in that entire range of salinities, earning the term euryhaline, and this would  obviously be impossible if they allowed their body fluids to equilibrate freely with the seawater. Aphanius survive because over this range the internal milieu of an Aphanius remains reasonably stable, mainly due to some neurohormonal trickery used to manage the escape of water through the body surfaces.

This sort of adaptation became necessary probably in the Ordovician and Silurian, when the primitive vasotocin gene underwent duplication in the jawless fish, splitting into oxytocin and vasopressin. These hormones were probably still doing some old reproductive thing, but then various animals started venturing into brackish or even fresh water, which was markedly hypotonic in comparison to the salt water they had previously enjoyed. Being surrounded by water makes you somewhat defenceless against its osmotic power, and these creatures had to develop a mechanism by which they could eliminate large amounts of hypotonic fluid in order to maintain their extracellular osmolality within a normal range. At this stage, selectively water-permeable urogenital tracts became a desirable design feature, and vasopressin became involved as the regulatory hormone (presumably because its precursor already had some involvement in the urogenital tract). In short, at some stage, some osmoconformer ended up shackled to a specific osmolality value by the development of these mechanisms, and since then we all have some kind of inbuilt homeostatic setpoint for our extracellular fluid osmolality, albeit different for each descendent species.

Returning to the point we misplaced some paragraphs above, the homeostatic setpoint for human osmolality is set to around 285 mOsmk/kg, but this is not decided logically according to some strict physicochemical principles, but is rather something of an evolutionary accident, a tribute to the half-remembered salinity of an ancient sea. It is therefore not surprising that this variable is similar among terrestrial vertebrates, but differs markedly across aquatic animal species. Most birds, reptiles and mammals fall within a range of 250-350 mOsm/Kg, close to the same value as humans, but invertebrates amphibians and aquatic animals fall somewhere along a very wide spectrum, where the extracellular osmolality is not identical to the surrounding water, but is obviously influenced by it. Just have a look at this helpful table from Takei (2000):

Clearly some of these must be simple osmoconformers (as you can see the osmolality of the hagfish blood is pretty similar to the osmolality of the surrounding seawater), but even among devoted osmoregulators there are rather vast differences: for example, see the gap between the coelacanth (931 mOsm/kg) and the toadfish (396 mOsm/kg), both saltwater species. Among invertebrates the situation is even more unruly, with some species of freshwater artheropods maintaining a massive 60:1 osmolality gradient between their extracellular fluid and the surround pond (McNamara & Freire, 2022). Haemolymph osmolality can range from something in the 700 mOsm/kg range in some species of freshwater decapod, down to 80 mOsm/kg in Hydrozoa and rotifers.

Those few readers who remain sane after finishing this extensive digression are referred to Timothy J Bradley's Animal Osmoregulation (2009). The bottom line is, we terrestrial vertebrates all share roughly the same homeostatic value for extracellular fluid osmolality because it was decided by some primitive amphibious ancestor, based on the salinity of the primordial ocean at the time of our invasion of the land. Our cellular machinery works perfectly within this range mainly because it was already working perfectly before we heaved ourselves up on our stubby lobe fins and waddled up the shore. We kept this value presumably because from an evolutionary standpoint it was easier to adapt existing teleost osmoregulatory mechanisms to maintain it, rather than changing every single macromolecule to adapt their function to some new osmolality value. 

Tonicity as the stimulus for water regulation

Most simply, vasopressin adjusts renal water excretion in response to changes in extracellular fluid tonicity. Tonicity goes up, vasopressin is secreted to stimulate water resorption and lower tonicity. Tonicity drops, and vasopressin secretion is decreased, allowing more water to be eliminated. Tonicity specifically is referred to here, instead of osmolality, even though the two are occasionally (incorrectly) used interchangeably in this context. Osmolality is the number of osmoles of solute per kilogram of solvent, whereas tonicity is the osmotic pressure between two compartments related to the difference in the concentration of "effective" osmoles between them, and it is the latter that affects water regulation, mainly because of the design of the hypothalamic osmosensors.

Hypothalamic osmosensors

These are two organs in the vague vicinity of the hypothalamus:

• Organum vasculosum lamina terminalis
• The subfornical organ

They are also sometimes referred to as "circumventricular organs", in the sense that they are found somewhere around the ventricles.  These small lumps of nervous tissue are not anatomically interesting, and the main thing you need to know about their structure is that they protrude outside the blood-brain barrier: i.e the capillaries that supply them are fenestrated and the molecular traffic there is free from the usual censorship imposed by astrocyte foot processes. 

The whole point of these areas is to sense changes in the concentration of solutes in the body fluids. To be precise, these sensors - though usually referred to as "osmosensors" - do not sense changes in osmolality per se, but rather changes in tonicity. The reason for this is their mechanism of action. These cells sense tonicity by detecting their own osmotic swelling or shrinkage. They are mechanosensitive, as demonstrated by some elegant experiments by Prager-Khoutorsky (2017). When these cells are harassed by a biologist with a pipette, the mechanical deformation to their shape is transduced into a change in their firing rate, such that cell swelling results in slower firing, and cell shrinkage results in a barrage of rapid discharges:

So, osmosensing is a purely mechanical process, and this explains why only changes in tonicity, and not total osmolality, are important. Ineffective osmoles will equilibrate between the inside and the outside of these cells, producing no change in their volume, and therefore no change in the rate of their firing. 

Regulatory responses to changes in tonicity

These mechanotransducers are extremely sensitive - probably more than conventional teaching suggests. A threshold of 1%, or a 3mOsm/kg change, is usually quoted by the textbooks, implying that the osmosensors are supposedly incapable of detecting changes smaller than this. This frequently repeated 1% value seems to come from the interpretation of a study by Robertson & Athar (1976) who measured the response of eighteen healthy subjects to the effects of fluid deprivation and hypertonic saline infusion. After about forty minutes of slow infusion, there was a detectable change in plasma vasopressin levels, which corresponded to the first detectable change in plasma osmolality (about 3 mOsm/kg). 

Robertson Shelton and Athar then went on to editorialise in Kidney International, describing the satisfying linearity of their findings. From the slope of this line, they found that for every 1 mOsm/kg increase in osmolality, the vasopressin secretion increased by 0.34 pg/ml; i.e. they appeared to acknowledge that there is probably some secretion occurring in response to very small changes in osmolality, but that it was probably impossible to measure these:

"a change in plasma osmolality of only 1% (2.9 mOsm/kg) would be expected to change plasma vasopressin by about 1 pg/ml, an amount large enough to be detected by some immunoassays"

To be fair, they were writing in the 1970s, and were lucky to be able to measure picograms per milliliter with their primitive instruments. Moreover, one must recall that we still measure the osmolality of a solution by freezing point depression, and the change from adding one whole mole of something to a litre of water results in a drop of only 1.86º K, i.e. the instrument would need to accurately detect a 0.00186º K change in freezing point temperature (two one-thousandths of a degree) in order to measure a 1 mOsm/kg change in osmolality.

In short, though osmosensors probably micromanage your vasopressin secretion even more tightly than textbooks might say, in practice we may never be able to prove this, because our scientific instruments are much less sensitive than our own organae. We suspect this to be true because a borderline-measurable drop in osmolality (2.7 mOsm/kg, or 1% change) results in a huge incremental change in vasopressin secretion and renal water handling - a drop of plasma vasopressin level from 2 pg/ml to 1pg/ml, and a drop in urine osmolality from 500 to 250 mOsm/kg (Bichet, 2011).

Physiological limits of osmoregulatory mechanisms

The range of possible responses to tonicity is obviously going to be constrained by what the pituitary can pump out, and what the kidneys can achieve when properly motivated. Of these, the latter is a more important consideration. Certainly from the size and fragility of the posterior pituitary we can conjecture that it is probably incapable of great secretory heroism, and there are situations where it can become exhausted (eg. the relative vasopressin insufficiency of sepsis), but vasopressin is a hugely potent molecule and you do not need a thyroid-sized gland to secrete enough of it to make a difference. The main constraints on the regulation of tonicity are imposed by the effector organs, specifically by the kidneys, as there is a limit to how dilute and how concentrated the urine can get.

First, let's look at what happens at low tonicity. You would want to dump dilute urine, and therefore you would want to have minimum antidiuretic hormone around so you can produce the most dilute urine imaginable. Of course the minimum possible concentration of vasopressin is zero. We don't really know whether it ever drops that low in normal physiology (or whether there is some minimum level beyond which it cannot fall) because of the abovementioned problems with measuring extremely small concentrations of it, but we can extrapolate a line of best fit from osmolality-secretion response curves, such as these ones from Thompson et al (1979):

You can see that the linear response crosses the abscissa at around 285 mOsm/kg.  Irrespective of whether the amount of vasopressin here is zero or just undetectably low, the renal response is still the same - hosing diuresis. The osmolality of this urine is 40-50 mOsm/kg, which is as dilute as you can get. This would perhaps vary from person to person depending on how much solute they were obliged to eliminate, urea and glucose being the main ones (in every person some urea always needs to leave the body). Theoretically a severely hypoglycaemic person with extremely depressed protein metabolism could approach a minimum urinary osmolality of close to 0 mOsm/kg, but in reality there is always some stray solutes in there, as demonstrated by the lower edges of this diagram from Baylis (1985):

 

As you can see from this plot, beyond a certain point increasing vasopressin levels do not produce a proportional increase in osmolality. Here again the kidneys are the limit ing factor. It is impossible to retain all the water, and some necessarily escapes, but the urine can be rendered maximally concentrated. Humans seem to max out at about 1200 mOsm/kg, or roughly the maximum osmolality of the deepest renal medulla where all the renal countercurrent concentrating effects are happening. It appears this medullo-urinary equlibration is achieved at a vasopressin concentration of something like 4-5 pmol/L:

So, it appears that at an extracellular tonicity of something like 290-300 mOsm/kg, vasopressin secretion reaches a point of maximal renal concentrating effect. But the reader will have noticed that the plasma vasopressin concentration graph extends way beyond this threshold. The pituitary just keeps secreting it, and the level keeps rising, long after the kidneys have maximised their response. This might seem to defy logic (why keep secreting it if it will have no further water-retaining effects), but it is in fact still having various valuable non-renal effects, such as its effects on V1 receptors such as vasoconstriction and blood flow redistribution. 

Association of vasopressin secretion with thirst

Thirst behaviour and vasopressin are not perfectly linked, but they obviously exist in close association, as you might expect - because once you have minimised renal water loss, the only other avenue of decreasing your tonicity is to increase your water intake. Bichet (2018) and Arai et al (2013) explore this association in way more detail than anybody could possibly ever need. The briefest possible version of describing these connections is this: 

• Osmosensor neurons, particularly those of the subfornical organ, project to the anterior cingulate gyrus which is responsible for the conscious perception of thirst, and which then leads to water-seeking and drinking behaviour
• These same neurons also project to the vasopressin-producing neurons of the hypothalamus
• Increased extracellular tonicity stimulates both, and the subjective sensation of thirst increases in its intensity in parallel to the rate of vasopressin secretion.
• This generates an association between vasopressin levels and thirst, but they are not causally connected, and a vasopressin infusion will not make your patient thirsty.
• Moreover, various descending behavioural and sensory pathways can control vasopressin release on the basis of thirst and satiety. For example the sensation of water on the tongue or the presence of dilute fluids in the gastrointestinal tract can preemptively decrease the secretion of vasopressin even before a fall in extracellular fluid tonicity occurs.

Baroregulation of vasopressin secretion

Vasopressin has famously weird vasoconstrictor effects and gets secreted in response to a loss of circulating volume, as a part of a stereotypical autonomic response. In fact it gets secreted as the result of that stereotypical response, which means it can be stimulated by basically anything that activates the sympathetic nervous system, including pain and emotional distress (a famous experiment from the 1950s described by Verney involved the investigators torturing a dog with electric shocks, and noting that the distress this caused increased the heart rate and blood pressure along with causing vasopressin release). The most important implication of this is of course SIADH, which can be produced by things like surgery; but for the CICM First Part Exam candidate, appropriate release of ADH (in response to hypovolemia) is more important, because it could easily form a part of an annoying viva.

Hypovolemia, or more precisely hypotension at the carotid baroreceptors, is the most potent nonosmotic stimulus for vasopressin release. This relationship is so reliable that Baylis (1987) actually described it as an equation, where  

log (Plasma vasopressin concentration) = 0.06 × (ΔMAP as a percentage + 0.67)

Experimental data suggests that the amount of vasopressin released in these cases is quite substantial. Values in excess of 500 pmol/L were recorded by Baylis & Robertson in 1979 when they attacked healthy volunteers with ganglionic blockers until their MAP dropped by 50%:

Yes, we keep seeing alternating pg/ml or pmol/L in these diagrams, but that does not matter as much as you might think because, helpfully, the molecular weight of vasopressin is about 1084 Daltons and so 1 pg/mL corresponds with 1.08 pmol/L vasopressin. Whatever the notation, its clear that vasopressin release increases five-hundred-fold when the normal circulatory system is challenged with a pressure drop.  In case your mind immediately goes to a clinical correlation, septic patients and other critically ill people don't tend to mount a response as strong as this, and Sharshar et al (2003) recorded values of only around 12-15 pg/ml in septic shock patients. Similarly 

the VASST study measured baseline vasopressin levels of around 3 pmol/L in septic shock patients, which increased to about 70-100 pmol/L when they commenced on an infusion of about 1.8 U/hr (Gordon & Russell, 2010). 

Threshold for baroreflex activation of vasopressin release

The reader may have noticed the fat grouping of dots at the low end of the blood pressure change range in that graph from Baylis & Robertson, through which you could not draw the same "line of best fit" if they were taken independently, out of the context of the rest of the data. This represents a sort of hesitance on the part of the baroreflex, which does not tend to pull the vasopressin trigger until quite a substantial blood loss/pressure drop are detected. 

How substantial is substantial? The venerable Kerry Brandis text refers to a 10% change in intravascular volume, and as we all know the current crop of CICM First Part examiners all passed their exams because of this book, this 10% value should be viewed as definitive, and we should expect to see it incorporated into marking criteria. You also find it in official College propaganda and a whole host of peer-reviewed papers.  It probably comes from animal studies, lovingly detailed by Thrasher, 1994. 

The threshold is definitely high in humans, as Goldsmith (1982) were able to drop the CVP of their healthy subjects by 6mmhg and generate a veritable storm of plasma catecholamines but no real change in vasopressin release. Animal data is conflicting but also supports a vaguely 10%-ish threshold, for example  Courneya et al (1989), for whose unfortunate rabbits the vasopressin release threshold lay somewhere between 10% and 20% of blood volume loss, or Miller et al (1979), whose rats were volume depleted precisely by 10%, and who demonstrated that the resulting increase in vasopressin levels was mediated at least in part by catecholamines. For Dunn et al (1973), 8% was the volume loss required to stimulate vasopressin release. 

Conflict between hypovolemia and hypotonia

It's not much of a conflict: hypovolemia wins. The baroreceptor stimulus to vasopressin release overrides any signal from the osmosensors by shifting the vasopressin response curve to the right, making the posterior pituitary tolerant of progressively lower and lower tonicities. The result is an abundance of vasopressin even though the extracellular tonicity may be desperately low. We see evidence of this in hypovolemic hyponatremia, where the patient is depleted of water as well as salt. As volume is restored, the influence of the baroreceptors on the tonicity response is removed,  vasopressin secretion is turned off, and a profound polyuria can suddenly occur, potentially producing a disastrously rapid self-correction of sodium.