The discussion in this chapter revolves around the physiological effects of losing 1 litre of total body water. However, we are rarely confronted with a situation where this occurs in isolation. More often than not, body water is lost together with electrolytes (by spilling on to the surgeon, or by leaking out as sweat, or by diuresis). Though the initial focus is on the movement of fluids and electrolytes, this digression-prone author then also detours into the physiological consequences of dehydration in a broader sense.
Clinical signs at various stages of dehydration
In the 1918 Croonian lecture, after complaining about the research disruptions brought about by his recent military service, Major Waler B. Cannon (the Higginson Professor of Physiology at Harvard) went on to discuss "records of men who have missed their way in desert regions". The references he used included W.J McGee's "Thirst in the Desert" (1898), which was a publication of enormous influence at the time. It was an account of the early US-Mexico boundary surveys (1854-56 and 1891-96), made available online by the University of Arizona (appropriately, given the subject matter).
"Thirst in the Desert" is a record of McGee's wanderings in those arid regions, in the company of scientists but otherwise in the absence of even most basic medical supplies, and in an era which preceded the widespread use of intravenous rehydration fluids. He made observations of the signs and symptoms of dehydration, progressing from mild to severe.
The early stages of dehydration (loss of 0-5% of body water) began with a dry mouth and a craving for fluid. A tachypnoea and tachycardia may be present, but they are not necessary features. McGee notes that there are few objective findings in this early stage of dehydration, and the subjective findings may be reported by the domesticated wuss "not inured to desert life" in the absence of actual dehydration.
Progressively, having lost more than 5-10% of their body water, the desert adventurer develops sunken eyes and loses his skin turgor. Saliva and bronchial mucus become thicker and more viscous. The tongue sticks to the teeth; the eyes are irritated by the loss of moisturising secretions, and vision becomes blurred. The ears also hurt and tinnitis develops. The voice becomes hoarse as the vocal cords lose their flexibility. A throbbing headache develops, which resembles meningism (and which is probably due to the stretching of dural structures). A disturbance in the level of consciousness may occur at this stage; in McGee's words "irascability arises, and companions quarrel and separate". Beyond this, frank hallucinations and delirium will occur, due to a combination of uraemia and electrolyte derangement
Having lost more than 10% of their body volume places one at a considerable risk of death. Severe dehydration (even uncoupled from heat stroke) may lead to irreversible multiorgan system failure even in the face of advanced modern critical care. Typically, this stage is characterised by circulatory collapse. Hyperviscosity leads to myocardial infarction, stroke, and DIC. Acute tubular necrosis and ischaemic hepatitis typically complete the picture. In the hyperthermic person the course may be complicated by rhabdomyolysis and seizures. Overall, in the desert an abandoned person - previously well hydrated - is approximately 24 hours away from the second stage, and likely no further than 72 hours from death.
Obvious examples to the contrary abound. McGee himself reports on the survival of Pablo Valencia, a Mexican horseman lost for eight days in the Arizona desert, in full 35°C sunlight, with only about 7-8L of water in his possession. During that time, he rode perhaps 35 miles, and walked or crept another 100-150. For some five days, he consumed his own urine, which may have been the single action which saved his life. When they found him, he had lost 25% of his body mass (of which a major portion was probably fat). "...his ribs ridged out as those of a starveling horse;... his lips had disappeared as if amputated, leaving low edges of blackened tissue;... his eyes were set to a winkless stare, with surrounding skin so contracted as to expose the conjunctiva, itself as black as the gums". He was babbling incoherently, and appeared to be almost completely deaf and blind; however in spite of this some basic mental functions were spared and Pablo never lost certain common-sense survival skills. McGee marvels that he "apparently squandered little vitality in those aimless movements that hasten and harden the end of the thirst-victim". In spite of obviously near-terminal dehydration, Pablo recovered completely - albeit with little recollection of the latter stages of his ordeal, and with his hair now prematurely grey.
In contrast to McGee's lyrical accounts, what follows is a fairly dry discussion of the electrolyte disturbances in dehydration.
What am I losing
Well, its fresh water. Pure H2O. The loss of 1 litre of water means the loss of 1.4% of total body weight.
Changes to the initial conditions: distribution of volume
Strictly speaking, the manner in which the fluid is lost will determine the physiological response, and the pattern of compartment distribution will vary depending on what else is lost together with the fluid. This very artificial scenario I present resembles a situation where somebody is left to slowly dehydrate in the absence of drinking water.
Such a scenario is not unthinkable, but one begins to wonder, what are the kidneys doing while this dehydration is happening? Is the glomerulus still filtering? Is the bladder filling up with concentrated urine?
And how is this water being lost, is it evaporating from the lining of the lungs or is there a dialysis machine sucking it away, or is this a patient with an open abdomen in the operating theatre, steaming under those hot lamps?
To take into account such complexity would be to confuse the simple question of what happens in response to dehydration.
In simpler terms, one can say that because water traffics easily among the body fluid compartments, the loss of water from any one compartment will be rapidly corrected by the inflow of water from the other compartment.
Put another way, water will equilibrate and the proportional distribution of water will remain the same even though there may be 1000mls less of it.
- 80ml will be lost from the intravascular space
- 250ml will be lost from the interstitial fluid
- 670 ml will be lost from the intracellular fluid
Change to the initial conditions: distribution of electrolytes
Common sense dictates that the proportion of electrolyte distribution among the compartments in this simplistic model should not be altered by the loss of some solvent. So, the sodium concentration in the extracellular fluid should rise from 140 to around 143 mmol/L.
However, again we come up against the limitations of this model. The sodium concentration will change also because the organism, dehydrated in this way, will make attempts to conserve water, and in the course of these attempts the electrolyte concentration will fluctuate.
The osmolarity of the compartments has risen by 6.8 mmol or so, well above the 1% change in extracellular tonicity which is required to trigger a release of vasopressin. (in fact its about 2.4%) Thus, the pituitary releases vasopressin and the rate of water reabsorption from the collecting duct is increased in proportion to the rise in osmolality.
The change in intravascular compartment volume is around 2%. This is not enough to warrant a baroreceptor response.
This is explained in detail elsewhere. Essentially, its the response to extra body water which occurs even if there is no response from the osmoreceptors and baroreceptors, and is purely due to the fact that intravascular protein dilution results in diminished water resorption from the proximal tubule. In this situation, the balance is in favour of water retention, as the peritubular capillary contains more concentrated blood and therefore the gradient for resorption is higher.
Is any of this supported by experimental evidence?
To some extent, yes. The loss of 1 litre of water however is not a very large amount, only about 1.4% of the body mass of a normal 70kg organism. Likely nobody has experimented with such mild dehydration. Weirdly, it seems it was historically easier to get ethics approval to study severe dehydration in man.
The most cited articles on this topic are from the mid-1970s, written by Costill at the Ball State University, Indiana. Costil and his two friends recruited eight volunteers and forced them to pedal on an exercise bike in the heat until they lost 2, 4 and 6% of their body weight. The observed loss was due to sweat, so it was not entirely relevant to the model of “pure” water loss. Interestingly, it only takes 1.5 hours of heavy exercise to lose 2% of your body mass.
Thus, Shirreffs et al from Aberdeen collected fifteen Scotsmen and starved them for 37 hours, torturing them with questionnaires and blood tests. The dehydration was considerable – the volunteers lost about 2.7% of their body mass, probably about 2.0-2.5 litres of water. As they dried out, the investigators collected measurements of their plasma sodium and chloride concentrations. The rate of sodium increase seems to be roughly in proportion to that predicted by the simplistic model above; the volunteers started off with a sodium of 142, and it increased to 147 by the end of the 36 hours.
Extremes of dehydration
There are few papers regarding the fluid and electrolyte shifts at the extremes of dehydration in humans. These papers are typically written about palliative care patients, and to pester them unnecessarily with blood tests is frequently unethical. In spite of this, some investigators have published on this topic.
Specifically I refer to Dr Douglas Bridge, a palliative physician from the Royal Perth Hospital. A presentation of his research entitled "The physiology of prolonged fluid restriction in dying patients" is pretty much the only source I have on this issue.
The physiology of extreme and ultimately terminal dehydration is linked to the physiology of starvation. We are reminded that metabolism of fuel produces water. One gram of metabolised protein yields 0.41 ml of water, one gram of carbohydrate yields 0.72 ml, and one gram of fat yields 1.07 ml. A starving organism will preferentially metabolise fat, and thus each 10kg of fat metabolised will yield 10.7 litres of water. Furthermore, as protein catabolism decreases, urea excretion will also decrease and thus urine output will fall to a minimum, sponsoring water conservation. Thus, the terminally ill can persist for a surprisingly long time, with fat metabolism as their only source of water.
Dr Bridge measured the serum sodium of a series of terminally ill patients in his care, and charted its rise. This, to my knowledge, is the only such graph. I reproduce it below in tribute to his work, but without his permission, as I have no way of contacting him (no email address is listed).
Hypernatremia and circulating volume depletion leading to cardiac arrest are the causes of death from dehydration discussed in this presentation. It is difficult to argue, given that some of the patients ended up with a serum sodium of over 180mmol/L. I can only guess at the effect this might have on excitable tissues.
Predictably, pink fleshy humans are not biology's greatest champions as far as dehydration is concerned. There are many organisms better than us at this task. For instance, ruminants can withstand dehydration involving up to 40% body mass loss. More so the anurans: the leopard frog Rana pipiens can dehydrate to (and safely rehydrate from) 50% of its body mass.
Or perhaps you would like to carry on metabolism without water altogether? No problem! Anhydrobiosis is the term. Specifically, one particular weird little larva can survive complete dehydration (i.e. zero molecules of water in the tissues, dry as the surrounding Nigerian rocks). In fact, you can submerge it into pure ethanol, and keep it there indefinitely; once the solvent evaporates the larva will cheerfully awaken. The key seems to be the synthesis of trehalose, a gelatinous disaccharide.
Anyway, I digress.
The respiratory system and dehydration
Much of the time when we discuss dehydration and loss of body water we tend to assume that the dehydrating process is primarily heat-related. Of course, this is often the case - but it is possible for a normothermic individual to lose vast amounts of body water in the presence of a totally normal ambient temperature. These matters are discussed in greater detail in the chapter on the heating and humidification functions of the airways. In summary:
- Normal airways heat and humidify the inspired gas mixture
- On expiration, some of the expired moisture is reclaimed by the upper airway
- Depending on respiratory rate and ambient temperature, that reclaimed fraction can be up to 70%
- The reclaimed fraction of water can be decreased by:
- Increased ambient pressure
- Decreased inspired gas humidity (eg. wall oxygen)
- Bypassing the upper airway structures (eg. ETT or tracheostomy)
- The total exhaled water content can increase in the following situations:
- Tachypnoea (moisture loss is proprotional to minute volume)
- Tachycardia ( moisture loss is proportional to cardiac output)
Circulatory consequences of dehydration
From the volume loss point of view, dehydration resembles haemorrhage, albeit typically occurring at a far slower pace. Haemorrhagic shock and the consequences of blood loss are discussed in greater detail elsewhere, and will be treated here only in brief. The response to hyovolemia is characterised by tachycardia initially, and ultimately also hypotension. The systemic vascular resistance increases and the perfusion of non-essential systems is sacrificed. In the case of Pablo, McGee's wandering desert horseman, no peripheral pulses were palpable (given that his hypovolaemia was very severe, having lost approximately 40% of his body water). All of this is well-known and boring. More interesting are the rheological changes of the blood.
The influence of dehydration on blood rheology
Dehydration results in haemoconcentration, and haemoconcentration predictably leads to increased blood viscosity, though the relationship is unpredictably non-linear A good article (Brun et al, 1995) describes the influence of dehydration on the blood viscosity of rugby players. The players all lost about 300-1000ml of water during some sort of "standardized submaximal exercise session"; haematocrit and plasma vicosity increased in a small and predictable manner, but whole blood viscosity increased a whopping 36%, which the authors attributed to increased red cell rigidity.
The influence of dehydration on the circulatory response to exercise
Instead of anaesthetised greyhounds and little baby chickens, in this civilised modern era scientists have discovered that the ideal physiological model of dehydration is the human athlete, similar to the rugbymen of the blood rheology study quoted above. Gonzales-Alonzo et al (1997) consented fifteen endurance-trained cyclists to exercise in heat without rehydration, until ~4% of their body weight had been lost. Armstrong et al (1997) selected ten university students to undergo a ninety-minute heat stress test, and then tortured them with rectal temperature probes and blood tests. The authors found that dehydration greatly affects circulatory workload: put plainly, the heart has to work harder to compensate for the decreased circulatory volume.
This is an interesting finding, given that oxygen-carrying capacity of haemoconcentrated blood is actually increased (in terms of grams of oxygen carried per liter of blood). The answer lies in the influence of the circulatory demands of exercise on the circulatory compensation for hypovolemia. Normal compensatory mechanisms rely heavily on the abandoment of skin and muscle vascular territories, emphasising the perfusion of vital viscera. In heavy exercise, muscle territories need to be well perfused in order to function optimally, and skin blood flow must increase in order for sweating to be effective as a method of heat transfer. Having lost their compensatory mechanisms, the dehydrated person is suddenly exposed to uncompensated hypovolemia. Stroke volume decreaseses (by 7-8% or about 11ml in Gonzales-Alonzo's cyclists). In this healthy cohort of people well adapted to exercise, the increase in heart rate was fortunately enough to prevent any decline in cardiac output (at least until hyperthermia was added to the dehydration).
Thermoregulatory consequences of dehydration
A good peripheral circulation is precisely what is required to maintain adequate thermoregulatory function under conditions of hyperthermia. And it is precisely what is not available to the dehydrated human. The normal compensatory response to hypovolemia is to concentrate perfusion in the core organs. The skin is sacrificed to save the brain and heart. Unfortunately with the sweat glands deprived of blood flow one is unable to create a satisfactory amount of convective heat loss. Furthermore, convection is only effective if the surface from whence the sweat is evaporating is well perfused, i.e. the circulation must carry hot blood to the skin surface where its heat can be exchanged with she sweat and carried away as vapour. A vasoconstricted skin remains dry and useless. The heat of the body is restricted in the core. This was demonstrated by Armstrong's dehydrated uni students. Their core temperatures were higher when they were exercising in a volume-depleted state, suggesting that their evaporative heat loss mechanisms were becoming impaired.
Neurological consequences of dehydration
McGee offers brilliant observations of the multifactorial delirium which develops in severe dehydration. "The disordered state of body and brain is often revealed by ceaseless talk", he complains. "The talk rambled on and on - all talking slowly, seriously, with appropriate look and gesture, not one consciously hearing a word".
Mental state changes in dehydration are best known not from Mexican cattle ranchers but from terminally ill cancer patients receiving palliative care. Dalal and Bruera (2004) bring this up as a discussion point regarding what precisely constitutes "comfort care". Delirium arises due to uraemia, hypernatremia, and decreased cerebral blood flow (though this branch of the circulatory system is the most jealously guarded). Hyperosmolarity and the associated rheological changes probably contribute to some extent by degrading the oxygen delivery capacity of the microcirculation. However, in the modern world, one frequently finds that the relationship is reversed: the patient is dehydrated because they are delirious, not the other way around.