This article appeared in the February 1996 Edition of the Catholic Medical Quarterly

Return to Feb 96 CMQ Page


Peter McCullagh. MD., DPhil., MRCP.

"The sensation of thirst is basic to our very existence. Its gratification is universally held to be one of the pleasures of life; it cannot be ignored, and if water be lacking, the sensation comes to dominate our thoughts and behaviour."1

"In physiologic terms, thirst is defined as a deep-seated sensation of a desire for water. When fully stimulated, it is one of the more powerful behavioural drives experienced by humans."2

"Dehydration is commonly seen in patients with impaired consciousness whose thirst mechanism is undamaged, but who are unable to ask for water."3


The question of whether provision of food and water is to be regarded as a basic nursing measure (which should be provided without question by any decent facility) or as medical treatment (provision of which is at the discretion of the medical attendants) has been extensively argued. Whilst the withdrawal of hydration and nutrition has been advocated as a general proposition for patients with a variety of conditions, much less attention has been directed to the specific capacities and limitations of individual patients. Apart from the likelihood of major variations within categories such as congenitally brain damaged infants, young adults with traumatically acquired injuries and elderly people with senile dementia, it defies logic to bracket individuals from such widely differing categories together as potential candidates for "withdrawal".

Debate on any subject can be substantially directed by the questions selected. Nevertheless, it is surprising that, given the effort devoted to arguing whether individuals with disabilities such as those identified above retain "personhood", so little attention has been given to the question whether any of the individuals from whom hydration and nutrition are to be withdrawn could possibly experience persistent thirst throughout the period from cessation of fluid provision until death. This is especially so given the attention rightly accorded infliction of the discomfort of thirst on members of non-human species. Consideration of the possibility of thirst experience seems to have been set aside on the basis that lack of indications from a patient of awareness of his or her surroundings automatically connotes lack of capacity for thirst. The predominance of assumption in reaching this conclusion is. most starkly exposed if one considers "locked in" syndromes in which retention of capacity to appreciate is accompanied by incapacity to communicate. On the infrequent occasions when the basis for the assumption that thirst cannot follow withdrawal of hydration has been examined, as for example in an article coauthored from the New York Society for the Right to Die', it appears to be very flimsy.

This paper summarizes some questions which tend not to be asked and indicates how responses already exist in the scientific literature. What is known about brain functions required for thirst experience? How soundly is that knowledge based? How confidently can knowledge about thirst sensation derived from animal studies be transferred to the human situation? How likely are the brain lesions of patients considered for withdrawal of hydration to be compatible with retention of the capacity for thirst? How can this likelihood be assessed in individual patients?

The central role, in thirst sensation, of a number of nuclei, or collections of nerve cells, in the hypothalamus, a region in the base of the cerebral hemispheres, has been appreciated for almost 50 years.5 The, anatomical structures responsible for mediating thirst were discussed extensively in a symposium held in honour of the physiologist E.F. Adolph in 1963. Many details have been added and amended since then but the basic outline remains. The structures involved in thirst sensation are grouped in the base of the cerebral hemispheres. The limbic forebrain structures (which include aggregations of neurons such as amygdala, hippocampus and septum) have been described as interacting in a neuronal circuit with the hypothalamus to mediate thirst and thirst- initiated behaviour. These basal parts of the cerebral hemispheres have been categorised as the "visceral brain" because of their central role in such fundamental activities as mating and nutrition. The "visceral brain" is recognised as the part of the cerebral hemispheres that has undergone the least change during evolution of mammals.6

Much of what is known about thirst mechanisms is not new. Surely it is reasonable that this information be introduced into public discussion of the pros and cons of the withdrawal of hydration and nutrition from patients who are in a stable, chronic state of impaired consciousness? Whilst a complete description of all the neurological pathways capable of mediating thirst is not yet possible, it already appeared by the time of the Adolph symposium in 1963 that an extensive lesion might be required to guarantee removal of capacity for thirst, because "alternate pathways through the system would be available and, schematically at least, the lateral hypothalamic - medical forebrain bundle region in close proximity with the descending fornix component would be the only single locus where a lesion could conceivably interrupt all major outputs from the system."7 This description of the neural pathways mediating thirst concluded that it was prob- able that relatively parallel neural circuits coursing through the limbic system, hypothalamus and midbrain underlie the mediation of the primary drives.

The Basis of Knowledge about Thirst Mechanisms

Studies of experimental animals have utilized a wide range of species (see end of article) and have taken the form of attempting to stimulate thirst (by electrical impulses or by chemicals) or of selectively damaging specific parts of the brain and observing the effect of this on drinking behaviour. Apart from these physiological studies of function, there have also been a number of anatomical studies in which specialized staining techniques have been employed to detect the site of action of hormones and neurotransmitter chemicals believed to be involved in the mediation of thirst. Finally, clinical reports of patients with disturbances of the thirst mechanism as a result of disease have provided confirmation of the applicability of animal studies to humans.

Functional Studies of Thirst In Animals

The extensive series of experiments in which thirst has been investigated in animals has been documented by Fitzsimons.8 Three general features of these studies require emphasis. The first is the wide range of species that has been used, including laboratory rodents, dogs, sheep, goats, cattle, primates and birds. The second is that many of these experiments were undertaken before animal welfare considerations led to increasing restrictions on the use of some species and are now most unlikely to be, repeated. To some extent, this explains why a great part of the information available concerning. thirst mechanisms was acquired several decades ago. The third striking feature of thirst studies is the high degree of concordance between conclusions drawn from observations of different species. Whilst there has been elaboration relating to the precise identity of some of the brain structures involved and to that of the hormones and neurotransmitters mediating function, the general conclusions drawn by Fitzsimons have not been seriously challenged.

From the early 1940s, a succession of investigators demonstrated that the application of electrical stimuli to an animal's hypothalamus (the most basal part of the cerebral hemispheres) could evoke drinking. Other experiments in which drinking was produced by the injection of minute volumes of concentrated saline into the same region supported observations on the effects of electrical stimulation and, furthermore, suggested a mechanism that could be directly responsible for normal stimulation of the "thirst centre", namely an increase in the osmolarity of body fluids. These results supported earlier research which had shown that production of the anti-diuretic hormone (ADH), which is responsible for stimulating the kidneys to conserve water, was regulated by the hypothalamus. It would seem reasonable for mechanisms which have co-evolved to maintain fluid balance to be mediated by structures that are anatomically approximated.

Experiments involving observations of the effects of destruction of specific minute areas within and around the hypothalamus on the capacity to experience thirst have refined understanding of the precise identity of the clusters of neurons constituting the thirst centre but have not altered the original conclusion that this function has a hypothalamic location. There continues to be uncertainty about whether more than one function (for example both the monitoring of body fluid osmolarity of concentration and the sensing of thirst) is collocated in one cluster of neurons or in two discrete, but very close, structures.

Observations of the Impact of Disease on the Capacity of the Human Subject to Experience Thirst

As indicated in the quotation above from Hays et al,3 the abundant information available on the identity of those brain structures responsible for mediating thirst does not suggest that impairment of conscious state would be likely automatically to remove the capacity to be thirsty. On the other hand, it would be expected that diseases involving the hypothalamus, or some parts of it, could remove the capacity to experience thirst. There have been reports of cases of individuals with disease confined to restricted areas in the hypothalamic region who have, in consequence, lost the ability to be thirsty. Reports of loss of thirst, or adipsia, have been rare but have nevertheless been most informative, especially when accurate identification of the lesioned part of the brain could be achieved by independent means. The report of Hays et al, described a patient who, in addition to adipsia, had other features suggestive of damage to the hypothalamus3. In a review of neurological abnormalities that could interfere with maintenance of normal fluid balance. Robertson et al. made the point that these could involve loss of the capacity to experience thirst or retention of this capacity together with impairment of the capacity to respond to thirst.2 As will be indicated below, it is likely that the groups of neurons responsible for these two functions, although both located in the hypothalamus, are sufficiently separated to allow selective damage to either one on its own. One of the most specific (and informative) instances of damage to the thirst centre was that reported by Conley et al. in which a complete inability to experience thirst accompanied retention of the capacity to release anti-diuretic hormone in response to any increase in the concentration of body fluids.9 A loss confined to one of these two functions suggest their location in separate groups of neurons. The precise identity of the thirst centre and its relationship to other structures in the hypothalamus does not affect the general conclusions that can be drawn about the capacity of brain-injured patients to experience thirst. It may assist in drawing some inferences about capacity for thirst in patients with hypothalamic damage.

Anatomical Studies of Neuronal Circuits Mediating Thirst

It has been possible to locate neurons that are responsible for producing, and are able to respond to, molecules mediating thirst sensation. The hormone angiotensin is involved in the regulation of body fluid maintenance. Its application to the subfornical organ (a structure separate from the hypothalamus but closely connected to it by nerve tracts) of rats, which were water-sated at the commencement of the experiment led to renewed drinking.10 Subsequently, receptors that had the specific capacity to bind angiotensin were demonstrated on the cells of the subfornical organ in cats, confirming the involvement of this structure in the neuronal connections responsible for thirst.11   More recently, angiotensin has been shown to stimulate thirst in most species. it is likely that angiotensin produced in the brain has a role in regulation of thirst.12

How Accurately Does Water-Seeking Behaviour by an Experimental Animal Indicate that Thirst is Experienced?

As most of the extensive research on the mediation of thirst and the identification of hypothalamic-related structures that are involved in this process has been undertaken in experimental animals, it is not possible to confirm by direct questioning that thirst is being experienced. Additionally, whilst it has been possible to validate the absence of thirst in the rare adipsic patient with otherwise intact brain function, it is not feasible directly to confirm the retention of thirst by individuals with an intact hypothalamus who are incapable of communication because of other brain damage or disease. Nevertheless, the strength of the similarities, across a wide range of species, between thirst mechanisms and the brain structures mediating them is such as to justify extrapolations from observations in non- human to human subjects (It is certainly comparable with the levels of inter-species similarity that are generally accepted as justifying extrapolation in relation to other physiological function).

That animals are actually experiencing thirst, when hypothalamic brain structures that have been linked to thirst are being stimulated, has been inferred from observation of their behaviour. As reviewed by Fitzsimons, electrical stimulation of hypothalamic centres in species as diverse as rats, dogs and pigeons led to "the usual complicated sequence of actions involved in the search for water and eventually terminated by the final consummatory act." He concluded that "Drinking and its motor accompaniments, and presumably the sensory ones as well, resembled the normal pattern of behaviour evoked by natural thirst stimuli."

It might be postulated, as a means of discounting the significance of animal studies for brain-damaged patients, that experimental stimulation of the hypothalamus evokes involuntary drinking by an animal that is not actually thirsty. The possibility, however becomes untenable in the light of an experiment reported by Andersson and Wyrwicka.13 Their aim was to determine whether stimulation of the "thirst centre" in the hypothalamus produced a real sensation of thirst. In preparation for the experiment, goats were trained so as to develop motor conditioned reactions in response to natural thirst. To this end, the animals were prepared by withholding water until thirst developed. They were then only supplied with water to the accompaniment of a complex conditioning stimulus. The individual motor patterns that developed in association with thirst and the provision of water were reinforced during the course of a fortnight of training. Each animal was then submitted to electrical stimulation of the hypothalamus and, in each case, the individual motor-conditioned reaction that had developed in association with "natural" thirst was displayed. This establishes a very strong case for the proposition that stimulation of the appropriate parts of the hypothalamus does induce thirst. The question that then becomes relevant to the circumstances of the severely brain-damaged and non- communicating patient is, what other brain function, especially on the part of higher centres, is required for induction of thirst.

Is a Functional Cerebral Cortex Required for Thirst?

The research summarized to this point indicates most strongly that normal functioning by some structures in the hypothalamus is essential if thirst is to be experienced and that experimental stimulation of those structures evokes thirst. It also appears highly likely that the water-seeking behavioural patterns elicited in animals by stimulation of the appropriate hypothalamic structures provide an accurate indication that the subject is actually experiencing thirst. The conclusion that an intact hypothalamus is required for thirst appears inescapably: the question is whether it is also sufficient. Phrased differently, what other parts of the brain must also be functioning normally if thirst is to occur?

"Two separate but related mechanisms are involved in this regulation: a perception of the need for water (thirst) and the physical action of acquiring it."14

It is necessary to draw a distinction between those parts of the brain, additional to the hypothalamus and its associated structures, which are required if thirst is to occur and those structures that will also he required in order that a motor response to thirst is to be apparent to an observer. It is apparent that the effector limb entailing peripheral nerves and motor neurons, together with any neuronal net- works required for the co-ordination of the complex of individual muscular contractions involved in the movement patterns characterizing water-seeking and drinking, would also have to be intact for thirst to be apparent to an observer. Furthermore, it is likely that some hypothalamic structures, additional to those responsible for perception of thirst, may be required. Proximate to the centres for perception of thirst, there appear to be other hypothalamic centres which are necessary for the actions of seeking water in response to thirst. Thus, experimentally induced lesions in the lateral hypothalamus of rats were observed to ablate the capacity to drink but not the wish to do so (rats which had been conditioned to press a lever when wishing to eat and drink continued to do so after placement of these lesions, but nevertheless, were unable to eat or drink).15 These observations accord with the proposal by Morgane that, at least in the rat, there are separate centres for feeding/drinking activity and for satiety. The experimental production of small lesions in different parts of the hypothalamus permitted dissociation of the drive to feed and drink from the responses to that drive.16 Consequently, a lesion at any point in the effector limb could leave a patient's thirst sensation intact while depriving an observer of awareness of it.

There are clinical indications that a normally functioning cerebral cortex is not required for thirst and some cogent experimental demonstrations that the cortex can be inactivated without impairing thirst. Fitzsimons drew attention to the occurrence of thirst in anencephalic infants lacking a developed cerebral cortex17 and also to the persistence of thirst and hunger in decorticate cats and sheep (animals in which connections between the cerebral cortex and the remainder of the brain had been experimentally severed).18

Whilst a functioning cortex is not required in order for experimental animals to experience thirst, there is some evidence that under abnormal circumstances, when hypothalamic thirst centres have been destroyed, the cortex may gradually take over some of the function of thirst perception which it does not normally mediate. It has been reported that rats, in which adipsia had been produced by destruction of the hypothalamic thirst centre, gradually recover. Subsequent intervention to depress cerebral cortical activity in these animals reverses this recovery: on the basis of this observation, it has been concluded "that cortical activity may facilitate and maintain recovery from lateral hypothalamic lesions by enhancing the activity of depressed but intact tissue adjacent to those lesions."19

The response to the question posed at the beginning of this section would appear clearly to be in the negative. Activity on the part of the cerebral cortex is not necessary for thirst.

Are There Likely to be Any Reliable Indicators of Adipsia in Patients with Impaired Consciousness following Major Loss of Brain Function?

It appears highly probable that integrity of the appropriate centres, or clusters of neurons, in the hypothalamus suffices to ensure the retention of capacity for thirst. Retention is unlikely to be affected by impairment of cerebral cortex function. Consequently, it would appear highly likely that thirst could be retained in the absence of any ability on the part of the subject to communicate this information. On the other hand, it is likely that there may be some indirect indications that hypothalamic structures concerned with thirst perception have been inactivated, should this occur.

Surgical interventions in experimental animals have had as their aim the production of extremely localized lesions in order to determine precisely whether disparate functions are mediated by discrete, albeit closely approximated, structures. In contrast, spontaneous lesions produced in human patients by tumours or vascular accidents are likely to be much less discriminatory and to involve adjacent structures mediating different functions. Some reports of human cases of adipsia have noted the impairment of other hypothalamic functions that can be independently and objectively measured: others have not done so.

An example of the former situation, in which deficits additional to adipsia have been produced is the case reported by Travis et al.14 This child manifested adipsia, but additionally appeared to have had some impairment of anti-diuretic hormone secretion, inadequate temperature regulation and polyphagia (excessive ingestion of food) with resultant obesity. As animal studies have demonstrated that the hypothalamic centres responsible for thirst and for osmoregulation (with control of antidiuretic hormone secretion) are discrete although closely apposed,20 and as the centres controlling body temperature and feeding are also discrete, it is evident that the lesion in this patient was not as localized as those experimentally produced. In another case, reported by Spiro and Jenkins, subarachnoid haemorrhage from an aneurysm produced a combination of adipsia and hypothermia whilst leaving the osmoreceptor (and capacity to secrete anti-diuretic hormone) and hunger centres intact.21

On the other hand, spontaneous lesions may also produce adipsia without affecting other, closely located, hypothalamic functions. Thus, in the patient of Conley et al., adipsia was not accompanied by any other evidence of hypothalamic impairment. On the basis of the, admittedly limited, well described clinical cases of hypothalainic disorders, it would appear reasonable to conclude that, in the absence of evidence of specific impairment of the function of other centres that can be independently measured, it should be assumed that the hypothalamic thirst centre remains intact and functioning in patients with severe brain injury or disease.

A countervailing position, namely that "there is some evidence for impaired thirst in the setting of advanced age or neurological impairment" has been advanced by Ahronheim and Gasner.4 Referring to withdrawal of hydration they also note that "the experience of observers is that this process is not painful." Leaving aside the conceptual difficulty of how one experiences on behalf of others, I believe that the primary sources which they cited as evidence for impairment of thirst have only minimal relevance to the question of capacity for thirst during total, ongoing fluid deprivation. Thus, one study demonstrated reduced (not absent) thirst in elderly patients deprived of water for 24 hours.22 Another study reported on a selected subgroup of elderly patients with reduced capacity for thirst which was associated with failure to ingest ad- equate quantities of fluid.23 There does not appear to be anything to suggest that these observations would be applicable to other than elderly patients or that they could be interpreted as indicating a continuous and complete loss of capacity for thirst in circumstances of total fluid deprivation.

Is It Practicable to Alleviate Thirst and Achieve Satiety by Means other than Replacing a Water Deficit?

The cessation of thirst is normally referred peripherally to the oropharyngeal region. Assuming that provision of fluids to a patient lacking capacity to communicate consequent upon injury or disease of the brain was to be discontinued, would there be an opportunity to satiate thirst by peripheral measures such as the maintenance of moisture in the mouth?

Investigation of dogs, in which an oesophageal fistula had been produced, revealed that satiety could be achieved in thirsty animals by allowing them to ingest water ad libitum. Although satiety was achieved, as indicated by voluntary discontinuation of drinking after ingestion of the volume required to restore body water balance, no restoration of water balance occurred as the ingested fluid was diverted through the fistula away from the stomach.24 Furthermore, satiety was only temporary, with thirst returning about an hour later.

Fitzsimons' review noted the early observations of Claude Bernard that moistening of the mouth failed to relieve thirst in dogs or horses with oesophageal or gastric fistulae which prevented the absorption of water.25 It also recalled that whilst dryness of the mouth had once been regarded as an essential component of thirst, this hypothesis had been discarded. Permanent dryness of the mouth following extirpation of the salivary glands had little effect on voluntary water intake with the exception of that associated with the ingestion of dry food ("prandial drinking").26 It is evident that, whereas dryness of the mouth may aggravate a sensation of thirst resulting from body water depletion, its alleviation will not remedy thirst in the absence of correction of the depletion. Nursing measures directed to maintaining oropharyngeal moisture levels will be without any tangible benefit in relieving thirst for a dehydrating patient.



There is a concordance between experiments in a range of animal species involving stimulation or destruction of accurately identified parts of the brain, which is supported by studies of neurotransmitter pathways and observations of patients, that the neurological centre for thirst is located in the hypothalamus. Whilst other pain of the brain are required for the behavioural patterns that occur in response to thirst, sensation of thirst can be demonstrated to persist despite very severe damage to other parts of the brain, for example decortication. Hypothalamic lesions can result in complete loss of thirst: other lesions are most unlikely to do so even though they may render an animal or patient incapable of responding to fluid deprivation in the expected manner. In the absence of independent demonstration of hypothalamic damage in the location of the thirst centre, any assumption that a patient with severe brain damage, from any cause, is adipsic cannot be justified.



  1. Fitzsimons. J.T. (1972) P469 in Thirst. Physiological Reviews, 52,468-561.
  2. Robertson, G.L., Aycinena. P. & Zerbe, R.L. (1982) Neurogenic disorders of osmoregulation. The American Journal of Medicine, 72, 339.
  3. Hays, R.M., McHugh, P.R. & Williams. H.E. (1963) Absence of thirst in association with hydrocephalus. The New England Journal of Medicine, 269, 227.
  4. Ahronbeim, J.C. & Gasner, M.R. (1990) The sloganism of starvation. The Lancet, 335,278.
  5. For a brief review, see Martin, J.B. & Relchlin, S. (1987) Clinical Neuroendocrinology, Davis, Philadelpia, P79 et seq.
  6. Morgane, P.J. (1964) Limbic-hypothalamic-midbrain interaction in thirst and thirst-motivated behaviour. P429 in Thirst, Ed. M.J. Wayner, Pergamon Press, Oxford.
  7. Fisher, A.E. & Cory, J.N. (1964) Chemical tracing of neural pathways mediating the thirst drive. P515 in footnote 5.
  8. See footnote 1, pages 500-515.
  9. Conley, J.B., Brocklebank, J.T., Taylor, I.T. & Robson, A.M. (1976) Recurrent hypenatremia: a proposed mechanism in a patient with absence of thirst and abnormal excretion of water. The Journal of Pediatrics, 89, 898.
  10. Simpson, J.B. & Routtenberg, A. (1973) Subfornical organ: site of drinking elecitation by angiotensin II. Science, 181, 1172.
  11. Phillips, 1. & Felix, D. (1976) Specific angiotensin II receptive neurons in the cat subfornical organ. Brain Research. 109, 531.
  12. Blair-West, J.R., Denton, D.A., McKinley, M.J. & Weisinger, R.S. (1992) Thirst and brain angiotensin in cattle. The American Journal of Physiology, 262, R204.
  13. Andersson, B. & Wyrwicka, W. (1957) The elicitation of drinking motor conditioned reaction by electrical stimulation of the hypothalamic "drinking area" in the goat. Acta Physiologica Scandinavica, 41, 194.
  14. Travis, L.B., Dodge, W.F., Waggener, J.D. & Kashemsant, C. (1967) Defective thirst mechanism secondary to a hypothalamic lesion: studies in a child with adipsia, polyphagia, obesity and persistent hyperosmality. The Journal of Pediatrics, 70, 915.
  15. Baillie P. & Morrison, S.D. (1963) The nature of the suppression of food intake by lateral hypothalamic lesions in rats. The Journal of Physiology, 165, 227.
  16. Morgane, P.M. (1961) Electrophysiological studies of feeding and satiety centres in the rat. The American Journal of Physiology,201, 838.
  17. See footnote 1, page 474.
  18. See footnote 1, page 511.
  19. Teitelbaum, P. & Cytawa, J. (1965) Spreading depression and recovery from lateral hypothalamic damage, Science, 147, 61.
  20. Thrasher, T.N., Jones, R.G., Kell, L.C., Brown, C.J. & Ramsey, D.J. (1980) Drinking and vasopressin release during vasopressin release during ventricular infusions of hypertonic solutions. The American Journal of Physiology, 238, R340.
  21. Spiro, S.G. & Jenkins, J.S. (1971) Adipsia and hypothermia after subarachnoid haemorrhage. British Medical Journal, 11, 411.
  22. Phillips, P.A., Rolls, B.J., Ledingham, J.G.G., Forsling, M.L., Morton, J.J., Orowe, M.J. & Woliner, L. (1984) Reduced thirst after water deprivation in healthy men. The New England Journal of Medicine, 311, 753.
  23. Miller, P.D. Krebs, R.A., Neal, B.J. & Mclntyre, D.O. (1982) Hypodipsia in geriatric patients. The American Journal of Medicine, 73, 354.
  24. Thrasher, T.N., Nistal-Herrera, J.F., Keil, L.C. & Ramsey, D.J. (1981) Satiety and inhibition of vasopressin secretion after drinking in dehydrated dogs. The American Journal of Physiology. 240, E394.
  25. See footnote 1, page 499.
  26. See footnote 1, page 503.
Dr. Peter McCullagh M.D., D.Phil., M.R.C.P. is Senior fellow at the John Curtin School of Medical Research, The Australian National University.