Osmoregulation in clinical disorders of thirst appreciation

Normal human cellular function depends on constant tonicity of the extracellular fluid. As water is constantly lost from the kidneys, lungs, skin and alimentary tract, even in conditions of antidiuresis, there is a need to replace fluid deficits by the maintenance of adequate water intake. In healthy man, water homeostasis is so accurately controlled that plasma osmolality is precisely maintained within a remarkably narrow range of 282–298 mOsm/kg. This exquisite control is achieved by the close integration of the antidiuretic action of vasopressin (AVP), which regulates water excretion, and the sensation of thirst, which governs water intake. Disturbances of the secretion or function of vasopressin, or of the regulation of thirst appreciation and drinking behaviour, can cause profound clinical abnormalities in sodium and water homeostasis. Since the development of sensitive radioimmunoassays for the measurement of vasopressin in plasma, the physiological characteristics of AVP release have been established, and abnormalities of control of AVP secretion and function in diseases of water homeostasis, such as diabetes insipidus and the syndrome of inappropriate antidiuresis (SIADH), have been well defined. Because the definition of the thirst has been the topic of some controversy, and because the measurement of the sensation of thirst has been regarded as subjective and inaccurate, clinical disorders characterized by abnormal thirst have received comparatively little research attention. In recent years however, the development of accurate, methods for the measurement of thirst has improved our understanding of the physiology of thirst and advanced our knowledge of how abnormalities of thirst appreciation can contribute to clinical disorders of water balance. In this review, the physiology of control of water intake will be summarized and the pathophysiology and clinical features of clinical disorders of thirst appreciation will be described. The physiology of osmoregulation has previously been reviewed in depth (6) but we shall give a brief introductory overview. Plasma osmolality is the most important physiological determinant of vasopressin secretion. Changes in plasma osmolality are detected by specialised magnocellular neurones which animal data locate in the circumventricular organs (10; 50; 85) in the anterior hypothalamus, where it is thought that fenestrations in the blood brain barrier allow access of plasma solutes to osmosensitive neural tissue (89). The identification of water channels, or aquaporins, in brain tissue (aquaporin 4) has challenged this traditional theory. Rnase protection studies with aquaporin 4 probes indicated that the brain was the principal site of expression, prompting speculation that aquaporins in cell membranes may be the mechanism by which osmoreceptor neurones obtain their information about extracellular tonicity (1). Furthermore, as in situ hybridization studies identified particularly strong signals from the paraventricular and supraoptic nuclei, which synthesize AVP (48), the concentration of water channels in these nuclei may indicate that they have osmoreceptor properties independent of those of the circumventricular organs (37; 13). Further research is required to define the role of aquaporins in osmoregulation, but their identification in the brain is extremely exciting. Within the circumventricular organs, the subfornical organ (SFO) and the organum vasculosum laminae terminalis (OVLT) are the sites of the vasopressin osmoreceptors. Animal studies suggest that the osmoreceptors which control thirst are sited in the AV3V region, anatomically discrete from the site of the vasopressin osmoreceptors (91). Data from a human case report of defective osmoregulation of thirst, with preservation of osmoregulated vasopressin release, (31) would provide corroborative clinical evidence for the existence of separate osmoreceptors for thirst and vasopressin release. Neural signals from the osmoreceptors are transmitted to the supraoptic and paraventricular nuclei, where vasopressin is synthesized, and from where neurones terminate in the posterior pituitary, the site of AVP release into the circulation. It is thought that similar neural impulses from the osmoreceptors governing thirst sensation terminate in the cerebral cortex where they initiate drinking, but the neuroanatomy is less well understood. Electrolytic ablation of the ventral portion of the nucleus medianus has been reported to produce both adipsia and hyperdipsia in the rat, however (22, b) which suggests that the nucleus medianus controls the passage of signals from the circumventricular organs to the cerebral cortex. Elevations in plasma osmolality therefore produce increases in plasma vasopressin concentrations, as shown in Figure 1. Regression analysis reveals the physiological relationship between plasma osmolality and plasma vasopressin to be linear and to be described by the equation pAVP = 0.43 (pOsm − 284.3), r = + 0.92, P < 0.001, where vasopressin is expressed as pmol/l (80). The slope of this line reflects the sensitivity of the osmoregulatory unit. The abscissal intercept of the line, at a mean of 284.3 mOsm/kg in healthy man, represents the ‘osmotic threshold’, at which secretion of vasopressin into the plasma begins. Although sensitive cytochemical assays can detect minute quantities of circulating vasopressin at osmolalities below 284.3 mOsm/kg (8), these low concentrations of vasopressin have little antidiuretic effect, as a hypotonic diuresis develops when plasma osmolality is lowered below this ‘set point’. The concept of an osmotic threshold for vasopressin release is therefore biologically sound. The third characteristic of this equation is the close relationship between plasma osmolality and plasma vasopressin, as indicated by the correlation coefficient relating the two variables. There are wide inter-individual variations in the relationship between plasma vasopressin and plasma osmolality, but the relationship is very reproducible within an individual, on repeated testing (82). Results of measurement of (a) plasma vasopressin (pAVP) and (b) thirst in response to hypertonic saline infusion in ten healthy volunteers. - - - assay limit of detection for vasopressin (0.3 pmol/l). —- regression lines described in text. The close relationship between plasma osmolality and plasma vasopressin is lost during the act of drinking, which causes rapid suppression of vasopressin secretion, before changes in plasma osmolality occur (24; 77; Figure 2). The fall in plasma vasopressin concentrations is almost instantaneous and approximates closely to the plasma half-life of the hormone (77), suggesting that the act of swallowing itself may switch off vasopressin secretion. It has been proposed that the mechanism which inhibits vasopressin secretion during swallowing is the activation of oropharyngeal stretch receptors which send inhibitory signals to the hypothalamus via an, as yet unidentified, neural pathway (66). (a) Plasma osmolality, (b) vasopressin and (c) thirst responses to hypertonic saline infusion, followed by drinking (○) or water deprivation (•) in healthy controls. Adapted from 77) with permission. The sensation of thirst is subjective and not easily quantifiable by scientific measurement. 62) defines thirst as ‘a conscious sensation of a need for water and a desire to drink’, and according to this analysis has measured the osmotic threshold for the onset of thirst to be around 294 mOsm/kg (58), which is approximately 10 mOsm/kg above the osmotic threshold for vasopressin release. However, this analysis suggests that thirst rarely occurs in situations where plasma osmolality lies within the normal physiological range. This is contrary to clinical experience and suggests that osmoregulated thirst plays no role in governing fluid intake in physiological situations. The development of simple visual analogue scales for the measurement of thirst has led to this traditional definition of thirst being challenged. Most visual analogue scales are uncalibrated, 10 cm long lines, upon which subjects are invited to place a cross between two extremes—for instance ‘not thirsty at all’ and ‘extremely thirsty’. As there is close concordance between the responses to questions covering a range of sensations, including ‘how thirsty are you?’, ‘how much do you desire water?’ and ‘how dry is your mouth?’, the first question is most commonly used. A number of groups have employed this simple technique to measure thirst and have shown in carefully conducted experiments that thirst appreciation does change within the physiological range of plasma osmolalities (63; ; 75) and that thirst ratings so obtained correlate closely with ambient plasma osmolality. Thirst responses defined by this method are highly reproducible within an individual (82) and correlate well with the subsequent volume of water drunk (75). Using a visual analogue scale, we have shown that the osmotic threshold for the onset of thirst is actually lower than that identified by Robertson and which is very similar to that for vasopressin release (Thompson wt al., 1986, Fig. 1). Calculation of a linear regression equation from the data shows that the relationship between pOsm and thirst is just as sensitive as that between plasma osmolality and plasma vasopressin. The osmotic threshold for thirst onset and the close relationship between plasma osmolality and thirst are highly reproducible within individuals on repeated testing (82). These data propose an alternative model of thirst appreciation, whereby thirst occurs when plasma osmolality rises above an osmotic threshold which is similar to that for vasopressin release, and the intensity of thirst varies according to the ambient plasma osmolality. In other words, within the physiological range of plasma osmolalities, thirst sensation is mild and initiates fluid intake sufficient to replace insensible water loss, whereas the more intense thirst generated by higher plasma osmolalities prompts the larger fluid intake required to replenish more severe fluid balance deficits. The non-osmotic switch-off of thirst during the act of drinking is also very similar to that of vasopressin (Fig. 2), with a rapid fall in thirst ratings before significant changes in plasma osmolality occur (77; 66), suggesting again that pharyngeal distension causes activation of neural inputs to the hypothalamus causing immediate suppression of thirst appreciation. This is clearly a sophisticated defence mechanism which protects against over-hydration. As there is a delay in gastrointestinal absorption of water, if drinking stopped only when plasma osmolality to there would be a significant volume of fluid in the or which when would lower plasma osmolality and a diuresis to water balance. The osmoreceptors for thirst are of blood is not in healthy or with has only properties Thirst may also be by or the are not well understood. There are data the of on There is evidence that thirst during hypertonic saline infusion but more recent data to of even at plasma concentrations, on thirst or vasopressin secretion Thirst is the most of diabetes insipidus and the to the fluid intake which is to replace In the of a normal thirst with would to to plasma tonicity and to by fluid intake. The regulation of thirst is normal in the of with diabetes insipidus Regression analysis of the relationship between thirst and plasma osmolality shows that the osmotic threshold for onset of thirst is similar to that of and that there is a similar close correlation between thirst ratings and ambient plasma osmolality The thirst mechanism in to and normal plasma osmolality, even The has diabetes insipidus as a of a in vasopressin in this animal model have shown that the is more to thirst than control The proposed that thirst in diabetes insipidus may in be to a which has been by to thirst, but not in the rat, however which would against an important role for in the of thirst in diabetes with diabetes insipidus are also in to healthy to blood volume but there have been no studies of thirst in human diabetes the thirst in diabetes insipidus to be an physiological response to of plasma osmolality. water drinking has been described in with In clinical the with the of thirst and vasopressin may be to from with clinical to the of diabetes insipidus with thirst is in to with of and thirst the whereas they in The of in depends upon the of AVP response to plasma osmolality during water deprivation or hypertonic saline with but is as of plasma osmolality does not drinking in the of antidiuresis may to the development of significant a is to control the to it is to even this increases the of if the is to to The of diabetes insipidus and has been described in with to anterior and a is in 1. In to are characterized by or of thirst appreciation, to vasopressin Most case studies have adipsia and vasopressin in response to osmotic but normal AVP responses to non-osmotic such as and Fig. which suggests that the site of the is the of this is with in the on Plasma vasopressin responses to hypertonic saline infusion (a) and (b) in a with diabetes normal range in healthy Adapted from with permission. Thirst is a of in diabetes and represents a physiological response to The characteristics of osmoregulated thirst and AVP release have been shown in studies to be normal in diabetes is not in diabetes which suggests that thirst in controlled diabetes is to volume to than the osmotic of is also normal in recent data has shown that of to the of water deprivation with thirst and drinking The thirst and drinking responses may be an important in the development of the which is characteristic of this is a of diseases in which a of thirst causes drinking and There have been which have been identified to have been with and these are in which is not with is to as water described by it was defined as an of fluid to Most of the in the from or no water drinking is diabetes first by which more accurately the clinical abnormalities with the The syndrome is characterized by water for which the the of thirst and which to the of hypotonic is and the only abnormalities are mild in plasma osmolality and plasma sodium The which is on of the in healthy but is in some with diabetes is clearly in with this syndrome Most with have been to have normal osmoregulation of vasopressin some have reported AVP release, which they of AVP release to It has been that vasopressin responses to may the in which is well in would suggest that it is more that the causes sensitivity to vasopressin, as do not normal AVP responses to water and osmolality does not of abnormalities of the control of thirst appreciation have been identified in clinical the osmotic threshold for thirst appreciation is lowered such that thirsty even at plasma osmolalities with of vasopressin secretion The fluid intake at low plasma osmolalities that vasopressin secretion is with the development of the hypotonic diuresis characteristic of the Because plasma vasopressin is it can be to from diabetes plasma osmolality to be low in and normal in diabetes of vasopressin secretion, such as water deprivation or hypertonic saline infusion are for the of the two The of thirst in is that more than healthy controls in response to a in plasma osmolality, and they do not thirst during drinking, and to thirsty even when of water It has been that is by abnormal function of the osmoreceptors governing thirst but the of drinking to thirst that non-osmotic control of thirst is also It more that there is a in the integration of osmotic and non-osmotic control of thirst in The between and has been for some of with have This from an desire for water independent of thirst, and represents an in the of drinking of to of and water may which can to of deficits and even occurs particularly if are with which water particularly and abnormalities of osmoregulation have been described in with AVP responses to osmotic have been reported in to with normal suppression of AVP drinking The osmotic threshold for thirst is lowered below that for AVP release, however, so that so as to plasma osmolality to with constant suppression of and therefore hypotonic diuresis is very to cause the which occurs in a significant of however, as the normal human is of to of water have been reported to have inappropriate secretion of vasopressin, particularly in with and there have also been of a of the for vasopressin release of which cause suppression of drinking is in such that they to thirsty drinking, and to have sufficient fluid to physiological The pathophysiology of the abnormal thirst in is but is the reported abnormalities in both osmotic and non-osmotic to be a in the integration of thirst in the higher than a in the osmoreceptors. thirst and drinking have also been reported in some with most is particularly for and is with to in with normal osmoregulation of AVP has been reported in a in was to have the thirst and also in with AVP release in a with In of with and or a range of abnormalities of osmoregulation with two and adipsia and thirst The data indicated that disorders of thirst are more than AVP in with and be in this The of of lies in the of have that can abnormal thirst and the to in this It is not however, the of osmoregulatory abnormalities is to the of or to the of the of cerebral with may of is is as plasma of osmolality does not thirst and there is a significant of the development of water and severe of the the but rarely the changes which are in to the drinking There are with the can for vasopressin is in response to non-osmotic such as or disorders are by inappropriate of thirst, with to in to Plasma osmolality and plasma sodium are and blood volume is with a blood plasma and thirst and no to is remarkably well but and are and the can to and do not fluid intake during such as or with rapid development of by adipsia or are but a number have been described in the for of abnormal osmoregulation of thirst and vasopressin release to be Figure 4 is a of the of abnormal osmoregulatory function that have been described. Plasma vasopressin and thirst responses to in plasma osmolality in of (a) (b) (c) from with permission. normal range of This is by an of the osmotic for both thirst and vasopressin release. do not thirst, or AVP their plasma osmolality has to their osmotic which is has been they and in to their osmoregulatory are water they thirst and AVP secretion, and a hypotonic The of and can to the of diabetes particularly if thirst is not This is to as and of the are normal and the is are from of by their to thirst and around their and so the only is mild will to to around of fluid this to be on a in of This is by thirst and vasopressin responses to osmotic as shown by the sensitivity or slope of the osmoregulatory lines for thirst and vasopressin release (Fig. The to some AVP has led to the that adipsia is to of the osmoreceptors however, has that as magnocellular neurones have osmosensitive properties of they are to detect and by vasopressin, even if the osmoreceptors are not adipsia has been described in with and of the and and sensitivity to vasopressin have AVP release and normal vasopressin release in response to and which the site of the at the The secretion of of AVP in response to osmotic and the sensitivity to the antidiuretic of AVP some of these the to and limit water The to to a and of vasopressin in response to plasma osmolality protects from with adipsia are also to during water and do not if they do As with depends on a fluid intake of around with for with to in the of particularly of the osmoreceptor is as osmoreceptor have of vasopressin release in response to plasma osmolality, and therefore as diabetes with of thirst to of the anterior of the of has been well to produce this syndrome occurs in the in approximately of on of the anterior and can be severe The blood to the the site of the osmoreceptor is from of the anterior cerebral and anterior and it is that these are at the of causes of osmoreceptor ablation and the site of the in adipsia is the than the magnocellular neurones which synthesize and vasopressin, is by the of these to normal of AVP in response to non-osmotic (Fig. It is from data that with adipsia can produce plasma concentrations of vasopressin than those required to concentration in response to or some of the that we have have also that they thirst when mean blood was lowered by approximately by the infusion of the to the activation of volume receptors Plasma vasopressin (pAVP) responses to in with diabetes insipidus with thirst (•) and adipsia normal range of - - - of detection of AVP assay (0.3 pmol/l). of osmoreceptor is the of some to AVP secretion. This as the in plasma of quantities of by during water The AVP is biologically as water is not and are at of if of are adipsia is the most to Because thirst and vasopressin are the are at of In some diabetes insipidus to of the anterior have significant of including loss, and are to their fluid intake is to the diabetes insipidus and be to approximately of fluid need and most at to detect changes in fluid balance the of this has been reported to thirst sensation in with diabetes insipidus and has also been to be but we have to be of little in our have been with in some The of a of water intake on and of plasma osmolality. This extremely of osmoregulatory as adipsia with osmoregulated AVP release, and has been reported in a case of a with delay and The of increases with particularly in individuals and those in In the of have a sodium which is with and and In of there is a of and in a at some during the of with of a at the of for does not to be the cause of the studies of the human have shown an in the of the and paraventricular nuclei the of years and of the nuclei of the magnocellular neurones which synthesize vasopressin These to those clinical studies which have shown AVP responses to hypertonic saline infusion and water deprivation in healthy with However, some have no in AVP responses to hypertonic saline infusion and lower AVP concentrations in subjects water deprivation have also been reported In to the osmotic secretion of AVP in the AVP secretion has been reported to be drinking does not to the oropharyngeal of AVP secretion in and the to AVP with in water to to AVP secretion may have been important in the of to an to have a plasma sodium of There are therefore a of abnormalities in the regulation of AVP secretion in the In the abnormalities in regulation of thirst are more Thirst and drinking responses are in in response to water deprivation or hypertonic saline infusion recent suggest that thirst responses to hypertonic saline infusion normal but the data has in This thirst is for during normal by the or normal vasopressin secretion, but if insensible water are by the to of water may well the of in The for the in thirst appreciation with is not may in as for instance in the syndrome of or a of cerebral may be to a in osmoreceptor In to the abnormalities of regulation of vasopressin secretion, only thirst has been reported to be abnormal in oropharyngeal of thirst is Although is not as a of thirst, the which is to the of in this therefore be to fluid and particularly in of and those with The syndrome of inappropriate antidiuresis is characterized by with inappropriate in plasma AVP and inappropriate are by definition as plasma osmolality is to below the osmotic threshold for thirst, the desire to be and drinking be In with have fluid balance which closely those of healthy individuals and do not evidence of suppression of There is no data on the regulation of thirst in of abnormal AVP secretion have been described have two with to of the during hypertonic saline in these there was preservation of the close relationship between plasma osmolality and thirst but regression analysis that the osmoregulatory line for thirst was to the with a lowered osmotic threshold for thirst (Fig. This would indicate that in the threshold for thirst is to the lower ambient plasma osmolality. The mechanism for this is not and (a) Plasma vasopressin (pAVP) and (b) thirst responses to hypertonic saline infusion in two with the normal in healthy Thirst and drinking are of normal osmoregulation in healthy of thirst appreciation, in have profound for water homeostasis. The of diabetes insipidus and can have particularly with the for Although the for thirst are subjective and their in clinical research has to our understanding of the physiology of thirst appreciation and the abnormal control of thirst in osmoregulatory The neural control of thirst appreciation and as a of for the of disorders of thirst have not yet been and of drinking the of The research by the which is in this was by from the and by close with