2.2 The neurohypophysis

The neurohypophysis is a complex neurohumoral system with a key role in body fluid homoeostasis and reproductive function. This chapter will concentrate on the physiology and pathophysiology of the two hormones made by the neurohypophysis, vasopressin and oxytocin, outlining the roles of both hormones together with the molecular, cellular, and anatomical basis of their regulation and function.

The neurohypophysis consists of three parts: the hypothalamic nucleii (supraoptic and paraventricular) containing the cell bodies of the magnocellular, neurosecretory neurons that synthesize and secrete vasopressin and oxytocin; the supraoptico-hypophyseal tract, which includes the axons of these neurons; and the posterior pituitary, where the axons terminate on capillaries of the inferior hypophyseal artery (Fig. 2.2.1).

The supraoptic nucleus (SON) is situated along the proximal part of the optic tract. It consists largely of the cell bodies of discrete vasopressinergic and oxytocic magnocellular neurosecretory neurons projecting to the posterior pituitary along the supraoptico-hypophyseal tract. In humans, vasopressinergic neurons are found in the ventral SON, with oxytocic neurons situated dorsally. The paraventricular nucleus (PVN) also contains discrete vasopressinergic and oxytocic magnocellular neurons projecting to the posterior pituitary along the supraoptico-hypophyseal tract. In humans, magnocellular neurons of the PVN synthesizing vasopressin are found centrally in the nucleus, with oxytocic neurons in the periphery. The PVN contains additional smaller parvicellular neurons projecting to the median eminence and additional extrahypothalamic areas including the forebrain, brainstem, and spinal cord. Some of these parvicellular neurons are vasopressinergic. Some vasopressinergic parvicellular neurons terminate in the hypophyseal-portal bed of the pituitary. These neurons cosecrete corticotropin-releasing hormone, and have a role in the regulation of adrenocorticotropin (ACTH) release.

The posterior pituitary receives an arterial blood supply from the inferior hypophyseal artery and the artery of the trabecula (a branch of the superior hypophyseal artery). Both these vessels derive from the internal carotid artery and its branches. The SON and PVN receive an arterial supply from the suprahypophyseal, anterior communicating, anterior cerebral, posterior communicating, and posterior cerebral arteries, via the circle of Willis. Venous drainage of the neurohypophysis is via the dural, cavernous and inferior petrosal sinuses.

Mammalian vasopressin is a basic nonapeptide, with a disulfide bridge between the cysteine residues at positions 1 and 6 (Fig. 2.2.2). Most mammals have the amino acid arginine at position 8. In the pig family, arginine is substituted by lysine. Oxytocin differs from vasopressin by only two amino acids—isoleucine for phenylalanine at position 3, and leucine for arginine at position 8. Nonmammalian species have a variety of peptides very similar to vasopressin and oxytocin. The similarities between vasopressin and oxytocin, and the degree of conservation among similar peptides across the animal kingdom, probably reflects derivation from a common ancestral gene.

The VP and OT genes lie in tandem array on chromosome 20, separated by 8 kb and 11 kb of DNA in humans and rats, respectively. Both genes are composed of three exons, and encode polypeptide precursors with a common modular structure: an N-terminal signal peptide, the specific vasopressin or oxytocin sequence, a hormone-specific mid-molecule peptide termed a neurophysin (Np), and a C-terminal peptide (Fig. 2.2.3). There is considerable homology between the Np sequences of the two genes, positions 10–74 being highly conserved at the amino acid level.

Hypothalamic-specific expression of VP and OT genes is conferred through selective repressor elements within both structural genes and the 5′ flanking sequences (1). Expression of the VP gene has been observed in extrahypothalamic tissues, such as adrenal gland, gonads, cerebellum, and probably the pituicytes of the posterior pituitary gland (2). Additional loci of control involved in VP expression in these tissues remain to be determined.

The regulation of VP gene expression is mediated through positive and negative regulatory elements in the proximal promoter. Several transcription factors bind to these elements; activating proteins 1 and 2 (AP1 and AP2) and cAMP-responsive element binding proteins (CREB) stimulate expression, while the glucocorticoid receptor negatively regulates expression (3, 4). The human, rat, and mouse OT promoters contain oestrogen-response elements and interleukin 6 (IL-6) response elements. However, the functional significance of these remain unclear (5).

VP gene expression can also be regulated at a post-transcriptional level. Water deprivation leads to an increase in length of the poly(A) tail of vasopressin mRNA, altering mRNA stability. Vasopressin mRNA processing may be further influenced through the interaction of a dendritic localization sequence, contained within the mRNA, with a multifunctional poly(A) binding protein (PABP). This RNA-protein interaction may play key role in RNA stabilization, initiation of translation, and translational silencing (6, 7).

Synthesis of vasopressin and oxytocin precursors occur separately in the cell bodies of specific magnocellular neurosecretory neurons of the SON and PVN. Generation of both mature hormones entails substantial post-translational modification of the large primary precursor. Following translation, the C-terminal domains are glycosylated and the precursors packaged in vesicles of the regulated secretory pathway which migrate along neuronal axons toward the nerve terminals of the neurohypophysis. Migration is microtubule dependent. During this process, the vasopressin and oxytocin precursors are cleaved by basic endopeptidases. The final products of processing, the mature hormone and the respective Nps, are stored as a complex in secretory granules within the nerve terminals of the posterior pituitary (8). An increase in the firing frequency of vasopressinergic and oxytocic neurons result in the opening of voltage-gated Ca^2? channels in the nerve terminals which, through transient Ca^2? influx, results in fusion of the neurosecretory granules with the nerve terminal membrane and release of their contents into the circulation. The hormone and its Np are cosecreted into the systemic circulation in equimolar quantities (9). Both Nps can bind both vasopressin and oxytocin in vitro. However, apart from acting as carrier proteins for vasopressin and oxytocin during axonal migration, Nps appear to serve no specific biological function.

The half-life of both vasopressin and oxytocin is short, that of vasopressin being 5–15 min (10). Both hormones circulate in the free form, unbound to plasma proteins. However, vasopressin does bind to specific receptors on platelets. Vasopressin concentrations in platelet-rich plasma are thus about fivefold higher than in platelet-depleted plasma (11). Both vasopressin and oxytocin are degraded by several endothelial and circulating endo- and aminopeptidases. A specific placental cysteine aminopeptidase degrades vasopressin and oxytocin rapidly during pregnancy and the immediate postpartum period.

Neurohypohyseal hormone release is modulated by sensory signals. In the case of vasopressin, the key sensory regulatory inputs reflect osmotic status and blood pressure/circulating volume. The relationships of the SON and PVN with the autonomic afferents and central nervous system nucleii responsible for osmo- and baroregulation are thus key to the physiological regulation of vasopressin. Functional osmoreceptors are situated in anterior circumventricular structures: the subfornicular organ, and the organum vasculosum of the lamina terminalis (OVLT) (12). Local fenestrations in the blood–brain barrier allow this neural tissue direct contact with the circulation. However, the presence of specific water channels (aquaporin 4) in both SON and PVN suggests that vasopressin neurons may have independent osmoreceptor function (13, 14). Moreover, vasopressin neurons of the SON and PVN express vasopressin receptors, highlighting the potential for autocontrol of vasopressin release through small branching neurites (15, 16). The act of drinking causes rapid suppression of vasopressin secretion. This response is mediated by oropharyngeal receptors. The afferent pathway(s) for this additional inhibitory influence on vasopressin release have not been identified.

Baroregulatory influences on vasopressin release derive from aortic arch, carotid sinus, cardiac atrial, and great vein afferents via cranial nerves IX and X. These project to the nucleus tractus solitarius (NTS) in the brainstem, from where further afferents project to the SON and PVN. Additional adrenergic afferents project to the SON and PVN from other brainstem nucleii, such as the locus coeruleus. Together, these act to integrate afferent inputs reflecting volume status. Interruption of ascending baroafferents increases plasma vasopressin concentrations, consistent with some degree of tonic inhibitory drive (17, 18).

The renin–angiotensin system is intricately involved in the regulation of vasopressin production. Circulating angiotensin II stimulates vasopressin secretion through receptors in the subfornicular organ and activation of subfornicular organ afferents. In addition, angiotensin II stimulates vasopressin release via a direct effect on vasopressin magnocellular neurons, where type 2 angiotensin II receptors have been identified (19). In rat, atrial natriuretic peptide (ANP) inhibits both osmo- and barostimulated vasopressin release via subfornicular organ afferents (20). The related brain natriuretic peptide (BNP) also inhibits vasopressinergic neurons in the SON in vitro. However, there are no data to suggest that ANP is a key regulator of physiological vasopressin release in humans (21).

Plasma osmolality is the most important determinant of vasopressin secretion. The osmoregulatory system for thirst and vasopressin secretion maintains plasma osmolality within the narrow limits of 284–295 mOsml/kg. The osmoregulation of vasopressin production and the physiological relationship between plasma osmolality and plasma vasopressin concentration is described by three characteristics: the linear relationship between plasma osmolality and plasma vasopressin concentration; the osmotic threshold or ‘set point’ for vasopressin release; and the sensitivity of the osmoregulatory mechanism.

Increases in plasma osmolality increase plasma vasopressin concentrations in a linear manner (Fig. 2.2.4). The abscissal intercept of this regression line, 284 mOsml/kg, indicates the mean ‘osmotic threshold’ for vasopressin release: the mean plasma osmolality above which plasma vasopressin starts to increase. Though there is no level of plasma osmolality below which vasopressin release is completely suppressed (22), such low levels of vasopressin have little antidiuretic effect. The concept of a threshold of vasopressin release thus remains a pragmatic means to characterize the physiology of osmoregulation; vasopressin release being increased from a basal rate by activation of stimulatory osmoreceptor afferents, and decreased to minimal values by removal of this drive and the activation of synergistic inhibitory afferents. The slope of the regression line reflects the sensitivity of osmoregulated vasopressin release. There are considerable interindividual variations in both threshold and sensitivity of vasopressin release. Twin studies indicate a strong heritable component in this variation. However, over time, these parameters are remarkably reproducible within an individual (23).

There are several physiological situations where the tight relationship between plasma osmolality and vasopressin concentration is lost. The act of drinking results in rapid suppression of vasopressin release, independent of changes in osmolality. In addition, the rate of change of plasma osmolality can influence the vasopressin response; rapid increases in plasma osmolality result in exaggerated vasopressin release. The osmotic threshold for vasopressin release is lowered in normal pregnancy, and a similar though smaller change occurs in the luteal phase of the menstrual cycle. Plasma vasopressin concentrations increase with age, together with enhanced vasopressin responses to osmotic stimulation. In contrast, thirst appreciation is blunted and fluid intake reduced. These changes, together with age-related decreases in renal handling of water loads and generation of maximal urine concentration, form the basis for the predisposition of elderly people to both hyper- and hyponatraemia (24).

As a principal determinant of fluid homoeostasis, vasopressin is a key player in maintaining haemodynamic integrity. Significant reduction in circulating volume stimulates vasopressin release through the activation of mechanoreceptors in the cardiac atria and central veins. Hypotension stimulates vasopressin release through the activation of aortic arch and carotid sinus afferents. In contrast to osmoregulated vasopressin release, progressive reduction in blood pressure produces an exponential increase in plasma vasopressin. Falls in arterial blood pressure of 5–10% are necessary to increase circulating vasopressin concentrations in humans. Changes in circulating volume and blood pressure trigger an autonomic and endocrine cascade resulting in a coordinated physiological response. Baroregulated vasopressin responses can be modified by other neurohumoral influences triggered as part of this coordinated response: ANP inhibiting and noradrenaline augmenting baroregulated vasopressin release. Importantly, baroregulated vasopressin release can occur at low levels of plasma osmolality—levels that would normally act to suppress vasopressin production. This apparent ‘hierarchy’ of regulation is important when considering the integrated physiological response to volume depletion and the pathophysiology of hyponatraemia.

Nausea and emesis are potent stimuli to vasopressin release, independent of osmotic and haemodynamic status. Manipulation of abdominal contents is another powerful stimulus to vasopressin release. Both contribute to the high plasma vasopressin values and consequent impairment of water load excretion observed after gastrointestinal surgery. Vasopressin release in response to these stimuli and others, such as neuroglycopenia, justify its classification as a stress response hormone.

Thirsts, and the drinking response to thirst, are key components maintaining fluid homoeostasis. The basis of thirst and the regulation of water ingestion involve complex, integrated neural and neurohumoral pathways. Animal data place the osmoreceptors regulating thirst in the circumventricular AV3V region of the hypothalamus, anatomically distinct from those mediating vasopressin release (25). Rostral projections to higher centres remain largely unmapped. In rat, lesions in the ventral nucleus medianus can produce adipsia and hyperdipsia, indicating this to be one route through which afferent pathways reach the cerebral cortex.

There is a linear relationship between thirst, determined by visual analogue scale, and plasma osmolalities in the physiological range (Fig. 2.2.5). The mean osmotic threshold for thirst perception is 281 mOsml/kg, similar to that for vasopressin release. Thirst occurs when plasma osmolality rises above this threshold, the intensity varying in relation to the ambient plasma osmolality. The functional characteristics of osmoregulated thirst, just as vasopressin release, remain consistent within an individual on repeated testing, despite wide variations between individuals (23).

As with osmoregulated vasopressin release, there are also specific physiological situations in which the relationship between plasma osmolality and thirst breaks down. The act of drinking reduces osmostimulated thirst, just as it does vasopressin release. There is a fall in the osmotic threshold for thirst in the luteal phase of the menstrual cycle. In contrast, thirst appreciation and fluid intake are blunted in elderly people. Thirst can be stimulated by extracellular volume depletion through volume sensitive cardiac autonomic afferents. In addition, hypovolaemia and hypotension lead to the generation of circulating and intracerebral angiotensin II, a powerful dipsogen (Table 2.2.1).

There are three vasopressin receptor (V-R) subtypes, encoded by different genes (Table 2.2.2). All have seven transmembrane spanning domains, and all are G-protein coupled. They differ in tissue distribution, signal transduction mechanisms, and function. There is 70–80% human–rat subtype homology at the amino acid level (26, 27). The human V2-R gene has been mapped to Xq28. The murine V2-R gene maps to a syntenic X-chromosome locus. In contrast to many other hormone receptors, the V2-R is up-regulated by its ligand.

Although vasopressin has multiple actions, its principal physiological effect is in the regulation of water reabsorption in the distal nephron. The hairpin structure and electrolyte transport processes of the nephron allow the kidney to both concentrate and dilute urine in response to the prevailing circulating vasopressin concentration. Active transport of solute out of the thick ascending loop of Henle generates an osmolar gradient in the renal interstitium which increases from renal cortex to inner medulla, a gradient through which distal parts of the nephron pass en route to the collecting system. This is the basis of the renal countercurrent osmolar exchange mechanism. The presence of selective water channel proteins (aquaporins) in the wall of the distal nephron allows reabsorption of water from the duct lumen along an osmotic gradient, and excretion of concentrated urine.

Thirteen different mammalian aquaporins have been identified to date. Seven (AQP1-4, AQP6-8) can be found in the kidney (28). Aquaporins act as passive pores for small substrates and are divided into two families: the water-only channels; and the aquaglyceroporins that can conduct other small molecules such as glycerol and urea. Specific structural arrangements within the primary, secondary, and tertiary structure convey the three functional characteristics of permeation, selectivity, and gating. The structure of aquaporins involves two tandem repeats, each formed from three transmembrane domains, together with two highly conserved loops containing the signature asparagine-proline-alanine (NPA) motif. All aquaporins form homotetramers in the cell membrane, providing four functionally independent pores with an additional central pore formed between the four monomers. Water can pass through all the four independent channels of water-permeable aquaporins, while the central pore may act as independent channel in some aquaporins (29, 30).

AQP1 is constitutively expressed in the apical and basolateral membranes of the proximal tubule and descending loop of Henle, where it facilitates isotonic fluid movement. Loss of function mutations of AQP1 in humans leads to defective renal water conservation (31). AQP3 and AQP4 are constitutively expressed on the basolateral membrane of collecting duct cells. They facilitate the movement of water from collecting duct cells into the interstitium. Expression of AQP3, but not AQP4 is modulated by vasopressin.

AQP2 is expressed on the luminal surface of collecting duct cells, and is responsible for water transport from the lumen of the nephron into collecting duct cells. Expression of AQP2 is vasopressin dependent; activation of the V2-R producing a biphasic increase in expression of the protein. Generation of intracellular cAMP by ligand activation of the V2-R triggers an intracellular phosphorylation cascade, ultimately resulting in the phosphorylation of nuclear CREB and expression of c-Fos. Activation of these transcription factors stimulates AQP2 gene expression through CRE and AP1 elements in the AQP2 promoter (32). In addition, vasopressin stimulates an immediate increase in AQP2 expression by accelerating trafficking of presynthesized protein from intracellular vesicles, and the assembly of functional water channels, composed of AQP2 tetramers, in luminal cell membranes (33).

Maximum diuresis occurs at plasma vasopressin concentrations of 0.5 pmol/l or less. As vasopressin levels rise, there is a sigmoid relationship between plasma vasopressin concentration and urine osmolality, with maximum urine concentration achieved at plasma vasopressin concentrations of 3–4 pmol/l (Fig 2.2.6). Following persistent vasopressin secretion, antidiuresis may diminish. Down-regulation of both V2-R function and AQP2 expression may be responsible for this escape phenomenon (34). Vasopressin has additional effects at other parts of the nephron; decreasing medullary blood flow, and stimulating an active urea transporter in the distal collecting duct. Vasopressin can also stimulate active sodium transport into the renal interstitium. These effects contribute to the generation and maintenance of a hypertonic medullary interstitium, thus increasing the osmotic gradient across collecting tubules, and augmenting the antidiuresis produced by the action of vasopressin on distal water channels.

Vasopressin is a potent pressor agent, its effects mediated via a specific membrane receptor (V1a-R). Systemic effects on arterial blood pressure are only apparent at high concentrations due to compensatory buffering haemodynamic mechanisms. Nevertheless, vasopressin is important in maintaining blood pressure in mild volume depletion. The most striking vascular effects of vasopressin are in the regulation of regional blood flow. The sensitivity of vascular smooth muscle to the pressor effects of vasopressin vary according to the vascular bed; vasoconstriction of splanchnic, hepatic, and renal vessels occurring at vasopressin concentrations close to the physiological range. Furthermore, there are differential pressor responses within a given vascular bed; selective effects on intrarenal vessels resulting in redistribution of renal blood flow from medulla to cortex. Such effects suggest that baroregulated vasopressin release constitutes one of the key physiological mediators of the integrated haemodynamic response to volume depletion.

Vasopressin is an ACTH secretagogue, acting through pituitary corticotroph-specific V3-Rs. Though the effect is weak in isolation, vasopressin and corticotropin-releasing factor act synergistically. Vasopressin and corticotropin-releasing factor colocalize in neurohypophyseal parvicellular neurons projecting to the median eminence and the neurohypophyseal portal blood supply of the anterior pituitary. Levels of both vasopressin and corticotropin-releasing factor in these neurons are inversely related to glucocorticoid levels, clearly suggesting a role in feedback regulation.

Vasopressinergic fibres and V-Rs are present in many areas of the brain, including the cerebral cortex and limbic system. These extensive neural networks are anatomically and functionally independent of the neurohypophysis (no neuronal connections being apparent, and the blood–brain barrier excluding the majority of circulating factors from these sites). In rodents, these central vasopressinergic systems have key roles in mediating complex social behaviour such as mating patterns. There are similar emerging data in humans, with association studies linking V1a-R gene sequence variation with autistic spectrum disorder, social phobia, and interpersonal behaviour patterns (35–37).

A number of other actions of vasopressin are listed in Table 2.2.3 (38–42).

The physiological regulation of water balance is intimately linked with that of circulating volume; common systems are involved in both processes. As sodium is the major cationic osmolyte, the interrelationships of sodium and water excretion with circulating volume regulation are key to appreciating the position of vasopressin in the physiology of fluid homoeostasis.

At plasma osmolalities of 285–295 mOsml/kg, osmolar balance can be maintained by vasopressin-dependent regulation of renal water loss. A rise in plasma osmolality within this range produces a progressive increase in plasma vasopressin to a concentration of 3–4 pmol/l, and antidiuresis. Further increases in plasma osmolality stimulate further vasopressin release, but this does not result in any further reduction of renal water excretion. Correction of plasma osmolality back to the range over which osmolar balance can be maintained by vasopressin requires thirst-stimulated drinking. As the osmolar threshold for thirst is similar to that for vasopressin release (284 mOsml/kg), the maintenance of water balance through a combination of osmoregulated vasopressin release and thirst is clearly a seamless, coordinated process of subtle complexity.

If excessive fluid volumes are consumed, greater than those demanded by thirst, plasma vasopressin levels are suppressed to below 0.3 pmol/l, resulting in maximum diuresis of up to 15–20 l/24 h. Ingestion of water in excess of this causes a reduction of plasma osmolality into the subnormal range, and hyponatraemia.

Vasopressin release is also regulated by other, nonosmotic stimuli. This complex regulation has a hierarchy, with significant physiological and pathophysiological sequelae. Hypovolaemia shifts the relationship of plasma osmolality with vasopressin concentration to the left. During moderate hypovolaemia, osmoregulation is maintained around a lower osmolar set point. As the degree of hypovolaemia progresses, baroregulated vasopressin release overrides the osmolar set point, and antidiuresis is maintained despite the potential for ensuing hyponatraemia. Coincident activation of the systemic and intracerebral renin–angiotensin systems stimulates drinking and augments vasopressin release, in addition to independent pressor and antinatriuretic effects. The homoeostatic response to hypovolaemia thus involves an integrated neurohumoral cascade, of which vasopressin is one component.

Oxytocin binds to specific G-protein coupled cell surface receptors (OT-Rs) on target cells to mediate a variety of effects concerned with reproductive function: the regulation of lactation, parturition, and reproductive behaviour. Recent data from animals lacking oxytocin because of targeted disruption of the oxytocin gene, have challenged this dogma, forcing a review of the physiological roles of the hormone (43).

In the rat, the stimulation of sensory afferents in the nipple by the act of suckling trigger a reflex cascade leading to synchronized firing bursts of oxytocic magnocellular neurons, and pulsatile oxytocin release corresponding to this burst activity. The released oxytocin acts on OT-Rs on smooth muscle cells lining the milk ducts of the breast, initiating milk ejection. Oxytocin is essential for completion of this milk ejection reflex in rodent. Mice lacking oxytocin fail to transfer milk to their suckling young, and this deficit is corrected by injection of oxytocin. In contrast, women lacking posterior pituitary function can breastfeed normally, illustrating that oxytocin is not necessary for lactation in humans. Pituitary lactotrophs express OT-R mRNA, and oxytocin released into the hypophyseal portal blood supply from the median eminence can stimulate prolactin release. However, the role of oxytocin in the physiology of prolactin release remains to be defined (5).

Oxytocin is the most powerful uterotonic agent identified. Furthermore, in many mammals there is both an increase in oxytocin secretion during parturition, and an increase in uterine responsiveness to oxytocin at term (5). These data suggest a key role for the hormone in the initiation and progression of labour. It is believed that falling progesterone concentrations toward the end of pregnancy lead to up-regulation of uterine myometrial OT-Rs, enhanced contractility, and increased sensitivity to circulating oxytocin. Stretching of the ‘birth canal’ during parturition leads to the stimulation of specific autonomic afferents, triggering increased burst firing of oxytocic magnocellular neurons and oxytocin release. A positive feedback loop is formed, with oxytocin both stimulating uterine contraction further and enhancing the production of local uterotonic mediators such as prostaglandins. It has been difficult to demonstrate increased circulating oxytocin levels in women during labour. This has been attributed to the difficulties of analysing pulsatile release, coupled with the short circulating half-life of the hormone due to the action of placental cysteine aminopeptidase. In mice lacking oxytocin, parturition is normal. Moreover, women with absent posterior pituitary function can have a normal labour. However, the importance of oxytocin in the birth process is highlighted by the effectiveness of oxytocin antagonists in the management of pre-term labour (44).

Recent data have highlighted an additional role of oxytocin in parturition. Maternal oxytocin produces a switch in fetal central nervous system (CNS) neurotransmission with enhanced inhibitory γ-aminobutyric acid (GABA)ergic signalling. This increases fetal neuronal resistance to hypoxaemic damage that may occur during delivery. These data suggest an adaptive mother–fetal signalling during parturition in which oxytocin is a major player (45).

OT-R expression is widespread in the CNS of many species. There is clear evidence that oxytocin has important influences on reproductive behaviour in rat; facilitating both lordosis and the development of maternal behaviour patterns (5). However, mice lacking oxytocin exhibit normal sexual and maternal behaviour, indicating these effects may be species-dependent. Central oxytotic transmission appears to reduce anxiety and hypothalamo-pituitary-adrenal stress responses in female rats (46). However, the same central oxytotic function may be required for normal adrenocorticotropin responses to stress. Together, these data suggest a complex role for oxytocin in the stress and other behavioural responses, with species and context-dependent differential effects (47, 48). Oxytocin release from both dendrites and nerve terminals of hypothalamic magnocellular neurons can be regulated by other neuropeptides, highlighting the potential for magnocellular oxytocin to integrate with central neurotransmission (49).

How are the proposed roles of oxytocin in reproductive function reconciled with both human and mouse data that highlight normal function in the absence of the hormone? First, there are interspecies differences in oxytocin-modulated processes that contribute important qualifications to the data. The mouse gravid uterus does not express OT-Rs, in contrast to human and rat. It is perhaps not surprising therefore, that parturition is normal in the oxytocin null-mouse. Similarly, in contrast to rat, maternal behaviour evolves gradually in mouse, and is not acquired rapidly in the postpartum period. Mouse may therefore not be a good model for the uterine and behavioural effects of oxytocin. Second, there is clearly variable redundancy in some of the physiological pathways in which oxytocin is involved. This redundancy may vary between species. The extrapolation of oxytocin’s role in normal physiology from those responses found in its absence should thus be made with caution.

There are no recognized clinical sequelae of oxytocin deficiency in humans. The pathophysiology of the neurohypophysis thus reflects the physiology of vasopressin and the regulation of water excretion. Defects in vasopressin production or action impact through disturbances in fluid and electrolyte balance. Another, less common, group of conditions reflect primary defects in thirst. In some cases, the two may coincide, reflecting the close anatomical and functional relationship of both processes.

Polyuria is defined by the excretion of urine in excess of 3 l/24 h (over 40 ml/kg per 24 h in adults and over 100 ml/kg per 24 h in infants). Diabetes insipidus is simply the excretion of large amounts of dilute urine. One of three mechanisms may be responsible:

Box 2.2.1 gives a classification of diabetes insipidus based on aetiology.

HDI (also known as neurogenic, central, or cranial diabetes insipidus) is due to deficient osmoregulated vasopressin secretion. In most cases it is a partial defect, with patients having inappropriately low plasma vasopressin concentrations with respect to concomitant plasma osmolalities. Presentation with HDI implies destruction or loss of function of more than 80% of vasopressinergic magnocellular neurons. Though persistent polyuria can lead to dehydration, given free access to water, most patients can maintain water balance through an intact thirst mechanism. HDI is rare, with an estimated prevalence of 1:25 000, and equal gender distribution.

Most cases of HDI are acquired. Improvements in imaging and an appreciation of the varied presentation of inflammatory/autoimmune forms are responsible for fewer cases being designated idiopathic. Trauma, either as a result of head injury or surgery, can produce HDI through damage to the hypothalamus, pituitary stalk, or posterior pituitary. Pituitary stalk trauma may lead to a triphasic disturbance in water balance; an immediate polyuria characteristic of HDI followed within days by a more prolonged period of antidiuresis suggestive of vasopressin excess. This second phase may last up to several weeks, and can be followed by reversion to HDI or recovery. Such a ‘triple response’ reflects initial magnocellular axonal damage; the subsequent unregulated release of large amounts of presynthesized vasopressin; and ultimately, either recovery or development of permanent HDI, as determined by the degree of initial neuropraxia/axonal shearing and damage. There is evidence that the polyuric phase may be associated with the presence of circulating inhibitors of vasopressin action, which may be partly processed vasopressin precursors (50). Not all phases of the response may be apparent. Recent data suggest acute HDI can occur in up to 22% of nonselected patients presenting with traumatic brain injury (TBI), persisting in some 7% of the total TBI cohort on long-term follow-up (51).

Although primary pituitary tumours rarely cause HDI, hypothalamic or pituitary metastases (for example, breast or bronchus) can present with HDI. In childhood, hypothalamic tumours, such as craniopharyngioma and germinoma/teratoma, are relatively common causes of HDI; together with developmental defects, such as septo-optic dysplasia (SOD), they account for up to 50% of cases in children (52). HDI can present in pregnancy, placental vasopressinase activity decompensating previously limited antidiuretic capacity through increased vasopressin degradation that cannot be matched by increased hormone release. This can revert to normal after delivery, though permanent HDI may ultimately develop if the natural history of the central defect is progressive.

Familial forms account for 5% of HDI. The Wolfram (WS) or DIDMOAD syndrome is a rare autosomal recessive, progressive neuro-degenerative disorder characterized by the association of HDI with diabetes mellitus, optic atrophy and bilateral sensorineural deafness. The natural history, of sequential development of the features, can be distorted by factors influencing presentation. Diabetes mellitus and optic atrophy are generally present in the first or second decade. HDI and deafness follow in the second or third decade. Additional features may then follow. Renal outflow tract dilatation is common, while gonadal atrophy and progressive ataxia with brain stem dysfunction can occur. Wolfram syndrome is caused by loss of function mutations in the WFSI gene on Ch.4p16. The gene encodes an 890 amino acid glycoprotein (wolframin). Non-inactivating mutations in the same gene are associated with autosomal dominant sensorineural hearing loss, suggesting the possibility of a spectrum disorder. An additional locus for Wolfram syndrome has been identified at Ch.4q22-24, suggesting genetic heterogeneity (53–55).

Autosomal dominant familial HDI is caused by mutations in the VP gene on chromosome 20. While it typically presents in childhood, the age of presentation varies considerably, reflecting variation in the progressive loss of vasopressin secretion. A variety of different missense and nonsense mutations within exons 1 and 2 of the VP gene have been identified in affected kindreds. Mutant vasopressin precursors accumulate in the endoplasmic reticulum of magnocellular neurons, to which they are neurotoxic. This explains the progressive loss of vasopressin release in the condition, and its dominant inheritance (56, 57). Growth failure may be an early clinical feature (58). The inherited HDI of the Brattleboro (BB) rat is due to a frame shift in exon 2 of the VP gene, resulting in a vasopressin precursor with an altered C-terminus which also accumulates in the endoplasmic reticulum of vasopressinergic neurons. Interestingly, the HDI of the BB rat is inherited in a recessive manner, in contrast to the equivalent condition in humans.

Circulating antibodies to vasopressin-secreting neurons can be found in 30% of patients classified previously as having HDI with no identifiable cause, implying an autoimmune aetiology. Presence of antivasopressin neuron antibodies in patients with HDI is particularly associated with pituitary stalk thickening on MRI. However, antivasopressin neuron antibodies can also be found at low prevalence in patients with HDI secondary to histiocytosis X and following pituitary surgery, suggesting the specificity of the test or the autoantibody response is low (59).

The strategy of investigation of HDI is to confirm the polyuric state, define its basis, and to explore possible primary aetiologies. After establishing significant polyuria of greater than 3 l/24 h in adults and excluding hyperglycaemia, hypokalaemia, hypercalcaemia, and significant renal insufficiency, attention should be focused on the vasopressin axis.

Direct measurement of plasma vasopressin in response to osmotic stimulation differentiates HDI from other causes of polyuria. However, access to reliable vasopressin assays is limited. Thus, a dynamic test using a surrogate endpoint of vasopressin release has been developed. This assesses the capacity to concentrate urine during the osmotic stress of controlled water deprivation: the water deprivation test. The period of water deprivation can be followed by evaluation of the antidiuretic response to exogenous vasopressin: the aim being to confirm renal sensitivity to vasopressin or establish renal resistance. A standard protocol is outlined in Box 2.2.2. HDI can be distinguished by urine osmolality less than 300 mOsml/kg, accompanied by plasma osmolality greater than 290 mOsml/kg after dehydration. Urine osmolality should rise above 750 mOsml/kg after desmopressin (DDAVP), indicating normal renal responsiveness. In contrast, failure to increase urine osmolality above 300 mOsml/kg after dehydration together with failure to respond to DDAVP is diagnostic of NDI. Patients with DDI should concentrate urine appropriately during dehydration, without significant rise in plasma osmolality.

In reality however, many patients have incomplete defects and manifest mild or moderate forms of diabetes insipidus. Moreover, prolonged polyuria of any type can impair urine concentrating ability through dissipation of the medullary interstitial concentration gradient, resulting in a partial functional NDI. The water deprivation test can be a poor discriminator in these circumstances. An accurate diagnosis of HDI can be made by direct measurement of plasma vasopressin during the controlled osmotic stress of a hypertonic 5% sodium chloride infusion (60). Patients with HDI have either undetectable vasopressin levels, or values falling to the right of the normogram relating plasma vasopressin to plasma osmolality. In NDI, plasma vasopressin is inappropriately high for the prevailing urine and plasma osmolality, indicating vasopressin resistance. In DDI, the relationship of plasma vasopressin to osmolality is normal. The test is not interpretable if the patient experiences nausea, a powerful non-osmotic stimulus of vasopressin release, during the test.

A pragmatic alternative to vasopressin measurements during hypertonic stress if there is diagnostic uncertainty following water deprivation is a controlled therapeutic trial of DDAVP: 10–20 μg of intranasal DDAVP per day for 2–4 weeks, with monitoring of plasma sodium every 2–3 days. Patients with DDI exhibit progressive dilutional hyponatraemia, whereas those with NDI remain unaffected. Patients with HDI experience improvement in polyuria and polydipsia, but remain normonatraemic.

Imaging of the hypothalamus, pituitary, and surrounding structures is essential in patients with HDI. MRI is the modality of choice. HDI is associated with the loss of the normal hyperintense signal of the posterior pituitary on T₁-weighted images. Signal intensity is correlated strongly with vasopressin content of the gland (61). As some hypothalamic germ cell tumours can be slow growing, imaging should be repeated at an interval of 6–18 months if the initial scan shows no demonstrable lesion. A negative scan at this stage should be taken as reassuring in the absence of a change in clinical features.

Patients with a urine output of less than 4 l/24 h can be managed by advising adequate fluid intake. The treatment of choice for those with more severe symptoms is the synthetic, long-acting vasopressin analogue DDAVP; given as an intranasal spray (5–100 μg daily), parenterally (0.1–2.0 μg daily), or orally (100–1000 μg daily) in divided doses. There is wide individual variation in the dose required to control symptoms. DDAVP has twice the antidiuretic potency of vasopressin, but has minimal vasopressor activity. It is well tolerated. Dilutional hyponatraemia is the most serious potential adverse effect. This can be avoided by omitting treatment on a regular basis (perhaps weekly), to allow a short period of breakthrough polyuria and thirst.

NDI is due to renal resistance to the antidiuretic effects of vasopressin. Primary familial forms are rare. X-linked recessive familial NDI is caused by inherited mutations of the V2-R gene on chromosome Xq28. Over seventy different mutations have been described: affecting receptor expression, ligand binding, and G-protein coupling. Most lead to complete loss of function; only a few are associated with a mild phenotype (62).

An autosomal recessive form is observed in 10% of kindreds with familial NDI, with normal V2-R function. Affected individuals harbour mutations of the AQP2 gene, leading to expression of dysfunctional water channels. Most mutations occur in the region coding for the transmembrane domain of the protein. Additional NDI kindreds harbour a mutation of the C-terminal intracellular tail of AQP2, leading to expression of a mutant protein that sequesters wild-type AQP2 (expressed by the normal allele) in nonfunctioning mixed tetramers in a dominant-negative manner (63). This form of NDI is inherited as an autosomal dominant trait.

More commonly, NDI is due to a variety of acquired metabolic or drug effects. The final common pathway producing NDI in many of these is down-regulation of AQP2 expression. NDI secondary to lithium toxicity can persist after drug withdrawal, and may not always be reversible.

Diagnosis of NDI is based on documenting inappropriately low urine osmolality with respect to circulating vasopressin levels, or lack of response to exogenous DDAVP. Secondary/acquired cases are managed by removing the underlying cause, and ensuring adequate hydration. Additional measures for persistent, severe symptoms rarely reduce urine volumes by more than 50%, though this may still be worthwhile. High dose DDAVP (4 μg IM twice daily) may produce a response in partial NDI, especially if the lesion is acquired. Thiazide diuretics (hydrochlorothiazide 25 mg/24 h), nonsteroidal anti-inflammatory drugs (ibuprofen 200 mg/24 h) and low-salt diets, singly or in combination, can also be effective. All probably work through reducing glomerular filtration rate, and interfering with the diluting capacity of the distal nephron.

DDI is a syndrome of excess fluid intake, and consequent polyuria. Though structural abnormalities may be the cause, more commonly it is a manifestation of primary hyperdipsia, psychiatric disease, or secondary to drug effects. DDI in the absence of other identifiable illness is compulsive water drinking. It is associated with abnormalities of thirst perception, including; a low osmotic threshold for thirst; an exaggerated thirst response to osmotic challenge; and an inability to suppress thirst at low osmolalities. The structural and/or functional basis for any of these abnormalities has not been identified. The association of DDI with affective disorders is well recognized. Up to 20% of patients with chronic schizophrenia have polydipsia. Although in some cases abnormal drinking is in response to beliefs founded in the primary thought disorder, complex abnormalities in both osmoregulated vasopressin release and osmoregulated thirst have been described. Whether these reflect long-term effects of drug therapy, or a primary defect in the central integration of thirst, is unclear.

Confirmation of the diagnosis of DDI is through direct or indirect demonstration of normal osmoregulated vasopressin release and antidiuretic action. As with many conditions, the treatment of DDI should address the underlying disorder. This can be difficult. Clozapine has been shown to reduce polydipsia in patients with refractory schizophrenia and a history of hyponatraemia. Whether this is due to an effect on central thirst mechanisms, or on suppressing disordered thought, remains to be clarified. Individuals with persistent DDI are at risk of hyponatraemia if treated with DDAVP, as fluid intake is maintained despite an obligate antidiuresis. In such cases a reduced fluid intake is the only rational treatment.

Hyponatraemia (serum sodium concentration less than 130 mmol/l) is common, occurring in about 15% of hospitalized patients (64). Hyponatraemia is not invariably associated with a low serum osmolality; high concentrations of other circulating osmolytes (for example, glucose), or a reduced plasma aqueous phase secondary to dyslipidaemia can result in hyponatraemia but normal plasma osmolality. Moreover, even when hyponatraemia is a true indicator of hypo-osmolality, it may reflect an appropriate physiological response. In order to maintain circulating volume in hypovolaemia, baroregulated vasopressin release proceeds despite plasma osmolalities well below the normal osmotic threshold. However, an individual with hypoosmolar plasma but a normal circulating volume, in whom the plasma vasopressin concentration is high for the prevailing osmolality, has a syndrome of inappropriate antidiuresis (SIAD) due to vasopressin excess. A variety of conditions are associated with SIAD, and to date four patterns of abnormal vasopressin secretion have been identified, as shown in Table 2.2.4 (65). Absolute plasma vasopressin concentrations may not be strikingly high; the key finding is that that they are inappropriate for the prevailing plasma osmolality. When this obligate antidiuresis is not accompanied by decreased water intake, haemodilution is inevitable.

Many conditions have been reported to cause SIAD (Box 2.2.3). SIAD is a nonmetastatic manifestation of small cell lung cancer and other malignancies. Some tumours are an ectopic source of vasopressin, and produce a type A syndrome. However, excessive posterior pituitary vasopressin secretion also occurs in association with malignancy. In fact the mechanism(s) of inappropriate vasopressin release in many cases of SIAD are not clear. The absence of an ectopic vasopressin source suggests a lesion in the neurohypophysis or its regulatory afferent pathways. The similarities between SIAD type B and the changes in vasopressin regulation in response to hypovolaemia and hypotension, suggest a single lesion in the baroregulatory afferent pathways. In contrast, the normal osmoregulated vasopressin release found in the type D syndrome suggests an increase in renal sensitivity to vasopressin, or the action of an as yet unidentified antidiuretic factor.

SIAD is a common mechanism of drug induced hyponatraemia. It can reflect direct stimulation of vasopressin release from the hypothalamus; indirect action on the hypothalamus via effects on higher centres; or aberrant resetting of the hypothalamic osmostat (66). Dopamine antagonists cause SIAD through stimulation of vasopressin release. Hyponatraemia is not restricted to one particular class of these agents, and has been reported with metoclopramide and newer antipsychotic compounds such as risperidone. Tricyclic antidepressants, monoamine oxidase inhibitors (MAOIs), and selective serotonin reuptake inhibitors (SSRIs) potentiate stimulatory central α₁ adrenergic input to vasopressin-producing neurons. Opiates also stimulate inappropriate vasopressin release through enhancing central adrenergic drive. SIAD is commonly associated with antiseizure medication. The frequency of hyponatraemia in patients treated with carbamazepine ranges from 4.8–40%, though the majority of such cases are asymptomatic. Carbamazepine increases both the sensitivity of central osmoreceptors, and renal sensitivity to vasopressin.

The major features in the diagnosis of SIAD are given in Box 2.2.4. The most frequent problem in clinical practice is distinguishing SIAD from chronic, mild hypovolaemia. In both conditions, urine osmolality tends to be higher than plasma osmolality. Plasma vasopressin will be detectable or elevated in both. Neither is therefore diagnostic of SIAD. The diagnosis hinges on confirming excretion of urine that is not maximally dilute in the context of a dilute plasma (i.e. urine concentration greater than 100mOsml/Kg). Renal sodium excretion should be above 20 mmol/l to make a diagnosis of SIAD. Below this value, volume depletion needs to be considered more likely. SIAD is often associated with urine sodium concentrations of 60 mmol/l or more. The hyponatraemia of chronic SIAD is not simply the result of haemodilution through reduced water excretion. SIAD is a volume-expanded state. Consistent with this, there is evidence of mild sodium loss as other regulators of volume homoeostasis attempt to minimize volume expansion (67).

Given the positive correlation between plasma vasopressin concentration and urinary excretion of AQP2, urine AQP2 excretion may be useful in the differentiation of SIAD from other causes of hyponatraemia (68). However, urinary AQP2 cannot differentiate clearly between hyponatraemic states associated with significant vasopressin production. SIAD and chronic hypovolaemia may generate similar plasma vasopressin concentrations and similar urine AQP2 levels and these two conditions are the most common differential diagnoses which we have difficulty in resolving. In addition, there are situations in which plasma vasopressin levels, urinary AQP2 excretion and renal concentrating ability are dissociated (e.g. following glucocorticoid replacement in hypopituitarism, central volume expansion, the newborn, the elderly). The clinical utility of the test thus remains to be clarified (69, 70).

The role of vasopressin production or action in producing hyponatraemia can be confirmed indirectly by assessing excretion of a standard water load over a fixed time: the water load test (Table 2.2.5). Normal subjects excrete 78–82% of the ingested water load in the 4h observation period. This is reduced to 30–40% in the presence of constitutive vasopressin production or action. The test is not essential to establish a diagnosis, although it can be helpful in planning management of chronic or recurrent hyponatraemia (60, 71).

Extreme endurance exercise is a physiological stressor. The magnitude of the physiological stress will reflect a number of factors: duration of the event; and the effort entailed. Non-osmoregulated vasopressin release is a feature of extreme endurance exercise, leading to a state of antidiuresis. If endurance athletes maintain a fluid intake in excess of water loss, hyponatraemia is inevitable. Athletes developing hyponatraemia demonstrate weight gain over the course of the event, consistent with water intake in excess of water loss. Health professionals attending endurance events need to be aware of the problem of exercise associated hyponatraemia. Athletes should be advised to follow their thirst. In addition, athletes who collapse during the course, or at the end of the event, should not be routinely resuscitated with large volumes of hypotonic fluid in the absence of appropriate indications and without biochemical monitoring as this may contribute to worsening hyponatraemia (72, 73).

While loss of function mutations of the V2-R are the cause of X-linked nephrogenic diabetes insipidus, rare individuals express the reciprocal problem: constitutively activating mutations in the V2-R that lead to vasopressin-independent, but V2-R mediated, antidiuresis resulting in persistent hyponatraemia. This nephrogenic syndrome of inappropriate antidiuresis (NSIAD) can have a variable phenotype. Although initially described in male infants with persistent hyponatraemia, the condition is not limited to males and may manifest in adulthood (74, 75). This is consistent with in the condition being X-linked but with variable expression in heterozygous females. Some 10% of patients with apparent SIAD have undetectable vasopressin. It is likely that at least some of these cases may be due to activating mutations of the V2-R.

Central salt wasting (CSW) is an acquired primary natriuresis found in a variety of neurological situations and a rare cause of hypovolaemic hyponatraemia. The underlying mechanism(s) involve increased release of natriuretic peptides and/or reduced sympathetic drive. The natural history of the process is key in establishing the diagnosis: hyponatraemia is preceded by natriuresis and diuresis with ensuing clinical and biochemical features of hypovolaemia. Depending on the point in the natural history at which the clinician meets the patient, urea and creatinine are generally elevated and there may be postural hypotension, in contrast to SIAD. The simple observation of weight loss over the period in question can be helpful. CSW is a particular concern in the neurosurgical patient: when autoregulation of cerebral blood flow is disturbed and small reductions in circulating volume can lead directly to reduced cerebral perfusion with secondary ischaemic brain injury. While both SIAD and CSW are associated with urine sodium concentrations greater than 40 mmol/l, the natriuresis of CSW is much more profound than that of SIAD and precedes the development of hyponatraemia. The management of CSW is volume replacement with 0.9% saline, balancing net sodium loss together with the requirement for circulating volume support (76).

The clinical impact of hyponatraemia secondary to SIAD reflects the combined effects of cerebral oedema and direct CNS dysfunction (Box 2.2.5). The clinical spectrum is wide. While values of serum sodium around 100 mmol/l are life-threatening, some patients with less marked hyponatraemia or in whom the problem has developed slowly commonly have mild symptoms or are asymptomatic due to CNS adaptation. CNS adaptation is limited: rapid changes in plasma sodium are accommodated less well than gradual changes, even if the scale of the change is relatively small. Moreover, this adaptation can complicate the management of hyponatraemia. Rapid correction of hyponatraemia following CNS adaptation can lead to significant changes in brain volume as the osmolar gradient across the blood-brain barrier alters. This may trigger CNS demyelination, a rare but serious complication of hyponatraemia and its treatment which develops within 1–4 days of rapid (>12 mmol/24 h) correction of plasma sodium. Other factors may play a role in susceptibility: concurrent hepatic dysfunction; potassium depletion; malnutrition; and it can occur even when sodium levels are corrected slowly Neurological manifestations include quadriplegia, ophthalmoplegia, pseudobulbar palsy and coma. Intervention to correct plasma sodium in SIAD must thus balance the morbidities of nonintervention with the risks of iatrogenic complications.

Chronic asymptomatic hyponatraemia with plasma sodium concentrations greater than 125 mmol/l, may not require specific treatment. More severe degrees of hyponatraemia, particularly if symptomatic, require some form of intervention. Correction of the underlying cause(s) is appropriate if the clinical situation allows it (treatment of infection, removal of the causative drug). Such approaches may prevent worsening hyponatraemia and allow the body’s own physiology to address the deficit in plasma sodium. Additional intervention should adhere to two key principles:

Fluid restriction of 0.5–1 l/day can be used safely when the clinical condition is not critical. The aim should be to have plasma sodium increase at a rate not exceeding 8–10 mmol/l per 24 h. Plasma sodium therefore needs to be measured regularly and all fluids need to be included in the restriction. Sodium intake should be maintained. It may be several days before sodium levels rise and it is important that a negative fluid balance is confirmed during this period. However, prolonged fluid restriction can be distressing and it is not always effective. The higher the baseline urine osmolality, the less likely fluid restriction is to work.

If the symptoms and signs of hyponatraemia due to SIAD are life-threatening, a more aggressive intervention may be required with hypertonic 3% sodium chloride. The aim of such an approach must be clear;

Clinical endpoints may be achieved through only a relatively small rise of 2–4 mmol/l in plasma sodium over 2–4 h. Importantly, normalization of plasma sodium is not the therapeutic target. Plasma sodium concentration should rise no more than 1–2 mmol/l per hour, with a total increment of no more than 8–10 mmol/l per 24 h. The volume of administered fluid required may be calculated through consideration of target plasma sodium, the sodium content of the administered fluid and the estimated deficit in plasma sodium based on body weight (64). If such an approach is used, it is imperative that the fluid regimen is reassessed at regular intervals, guided by careful clinical assessment and laboratory monitoring. However, the clinical utility of fixed replacement models in day-to-day practice is limited, especially if partial correction of plasma sodium to clinical endpoints is accepted and asymmetric increases biased toward more rapid changes in the first 1–4 hours of intervention are employed. An alternative approach is to use 100 ml boluses of 3% sodium chloride, with careful clinical and biochemical monitoring. Hypertonic fluid should be stopped when the defined clinical target or a sodium concentration of 125 mmol/l is reached, whichever is first. As before, the approach aims to reduce the neurological morbidity of hyponatraemia while minimizing the risk of precipitating osmotic demyelination (77).

If hyponatraemia persists or recurs after initial intervention, the underlying diagnosis should be reviewed and the basis for intervention reconsidered. If the diagnosis of SIAD remains intact, clinicians need to balance the merits of further incremental intervention with those of tolerating persisting hyponatraemia.

Demeclocycline produces a form of NDI and so increases renal water loss even in the presence of vasopressin. It is effective in treating hyponatraemia of SIAD at 600–1200 mg/day in divided doses. There is a lag time of some 3–4 days in onset of action. Treatment should be stopped if significant renal impairment develops. Lithium has similar effects and can be used as an alternative. However, the effects of lithium are less and the drug is associated with more adverse effects.

Urea increases renal free water excretion and decreases urinary sodium excretion. It can be used to treat the hyponatraemia of SIAD at doses of 30 g/day by mouth. It may be have clinical utility as an adjunctive therapy to allow reduction in water restriction and improvement in quality of life.

The nonpeptide V2-R antagonists (the vaptans) are a rational approach to the treatment of SIAD. They are aquaretic: increasing renal water excretion with no significant impact on renal electrolyte loss. Vaptans are either selective (V2-R specific) or nonselective (V2-and V1a antagonism). Both improve hyponatraemia associated with normal or increased plasma volume. Changes in plasma sodium can be seen within 4–6 hours. The drugs appear to be well tolerated. The developing role of vaptans in the management of SIAD needs to balance time course of action, tolerance, and long-term efficacy in specific clinical contexts (78, 79).

Adipsic and hypodipsic disorders are characterized by inadequate spontaneous fluid intake due to a primary defect in osmoregulated thirst. Patients are hypovolaemic and dehydrated, with elevated plasma sodium and urea. Despite this, they deny thirst and do not drink. If the defect is mild, the resultant hypernatraemia is often well tolerated. Severe disorders leading to marked electrolyte disturbances are tolerated poorly, and can lead to somnolence, seizures, coma and renal failure. Because of the close anatomical relationship of the osmregulatory centres for thirst and vasopressin release, adipsic syndromes are often associated with defects in osmoregulated vasopressin release and HDI, which can exacerbate electrolyte and water balance problems.

Four distinct patterns of osmoregulated thirst and associated vasopressin release are recognized (23). These are outlined in Table 2.2.6 and Figure 2.2.7. In addition, conditions producing adipsia/hypodipsia syndromes are outlined in Box 2.2.6.

The type A syndrome can be mistaken for HDI, as patients are hyperosmolar with a dilute urine. Formal assessment of thirst, by analogue scale during osmotic stimulation, confirms the diagnosis. Normal vasopressin responses to nonosmotic stimuli place the lesion responsible at the level of the osmoreceptor, rather than the vasopressin magnocellular neuron. The nature of the lesion remains unknown. Imaging is generally normal. Patients effectively osmoregulate around a higher osmolar set point, and are protected from extreme hypernatraemia, as are those with the type B syndrome. In contrast, type C adipsia is associated with complete lack of osmoregulated thirst and vasopressin release, consistent with complete destruction of osmoreceptors. Patients present with adipsic HDI. Specific precipitants include rupture and repair of anterior communicating artery aneurysm. One of the putative locations of the osmoreceptors mediating both thirst and vasopressin release, the OVLT, receives its blood supply from perforating branches of the anterior cerebral artery and anterior communicating artery. Some patients with the type C syndrome have persistent, constitutive low level vasopressin release. The resultant low level obligatory antidiuresis places such individuals at risk of dilutional hyponatraemia if large volumes of fluid are administered. Impaired osmoregulated thirst with normal osmoregulated vasopressin release (type D adipsia) is very rare.

Because patients with type A and type B adipsia are protected from extreme hypernatraemia, treatment is to recommend an obligate fluid intake of about 2 l/24 h, with appropriate adjustment for climate and season. If fluid balance cannot be maintained during intercurrent illness, in-hospital management may be required. The adipsic HDI of the type C syndrome can be difficult to manage. Structural and vascular causes of the type C syndrome may lead to associated cognitive defects in short term memory and task organization, which can complicate any intervention. The principle of management is to define an acceptable urine output (1–2 l/24 h) with regular DDAVP, and to vary the daily fluid intake depending on day-to-day fluctuation from a target weight at which the patient is euvolaemic and normonatraemic:

Daily fluid intake in litre = 1–2 l (i.e. the targeted urine output as dictated by the DDAVP dose set and taking into account insensible loss) + (target weight - daily weight in kg).

This formula, together with weekly checks of plasma sodium to avoid the creeping development of hyper- and hyponatraemia, can result in stable fluid balance (80).