Hypodipsic hypernatremia in eight dogs

Zusammenfassung Ziel: Beschreibung der klinischen und labordiagnostischen Befunde von Hunden mit einer durch eine Hypodypsie bedingten Hypernatriämie. Methode: Krankenberichte von acht Hunden mit einer hypodypsischen Hypernatriämie, diagnostiziert zwischen März 1997 und August 2002, wurden retrospektiv ausgewertet. Ergebnisse: Das mediane Alter zum Zeitpunkt der Vorstellung betrug 8 Monate (Spannweite: 4–27 Monate). Bei vier Hunden handelte es sich um Staffordshire Bull Terrier. Als häufigste klinische Sym-ptome traten Lethargie und Bewusstseinsveränderung auf. Der Median der höchsten gemessenen Natriumkonzentration jedes Hundes lag bei 196 mmol/l (Spannweite: 175–204 mmol/l). Magnetresonanztomographische Veränderungen fanden sich bei 2/4 Hunden (Hydrozephalus, fehlendes Corpus callosum). Sieben Hunde wurden hospitalisiert und mit intravenöser Flüssigkeitszufuhr für median 7 Tage (Spannbreite 1–15 Tage) behandelt; einem Hund wurde nur Flüssigkeit oral verabreicht. Zwei Hunde wurden während der Hospitalisation euthanasiert (einer wegen einer Pneumonie, einer aufgrund eines zerebralen Ödems) und fünf Hunde wurden entlassen und zu Hause mit zusätzlicher Gabe von Wasser im Futter behandelt. Klinische Relevanz: Bei hochgradiger Hypernatriämie stellt Hypodypsie eine seltene, aber wichtige Differenzialdiagnose dar. Die Ursache in den vorgestellten Fällen ist wahrscheinlich eine Dysfunktion der hypothalamischen Osmorezeptoren.


Introduction
Hypernatremia can be caused by loss of hypotonic fluids, gain of solutes or a pure water deficit (5). Pure water deficits are an uncommon cause of hypernatremia, since the body is very sensitive to changes in osmolality. An increase of 2% in osmolality maximally stimulates antidiuretic hormone (ADH) secretion, reducing the loss of free water (21). Even in the absence of ADH, as in individuals with central diabetes insipidus, the body is protected against the development of hypernatremia since an increase in osmolality also stimulates the thirst centre, resulting in increased drinking behavior. Thus, clinically significant hypernatremia is an uncommon finding in individuals that have an intact thirst response and free access to water (21). However, in rare cases the thirst response may be absent, resulting in severe hypernatremia. This syndrome of primary hypodipsia or adipsia has been reported in several dogs (1,3,7,11,12,15,14,24,25). Hypodipsic hypernatremia is also well recognized in humans (2,20) and has been reported in cats with intracranial disease (8,18). In the dog, congenital forms of the syndrome may be associated with a structural abnormality such as holoprosencephaly (24) or may be due to functional impairment without gross abnormalities. Acquired adipsia has been reported in a dog with granulomatous meningoencephalitis (17). A previous study has suggested that a dysfunc-tion of the hypothalamic osmoreceptors is responsible in some cases (7).
In this study we document the clinical and clinicopathological findings and the outcomes in a series of dogs with hypodipsic hypernatremia.

Materials and methods
Clinical records from the Queen Mother Hospital, Royal Veterinary College between March 1997 and August 2002 were reviewed retrospectively. Dogs with severe hypernatremia (serum sodium > 160 mmol/l; reference range 140-150 mmol/l) measured with an ion-selective-electrode (ISE) electrolyte system, historical signs of hypodipsia and no other identified cause of hypernatremia were included. The records were reviewed for details of the dog's signalment, clinical and clinicopathological findings, treatment and outcome.

Signalment
Eight dogs fulfilled the criteria for inclusion in the study (Table 1). All the dogs were purebreds with four (50%) being Staffordshire bull terriers. Although detailed pedigrees were not available, the Staffordshire bull terriers had all been obtained from separate breeding kennels. The median age at presentation was 8 months (range 4-37 months). Three dogs were intact females, three intact males and two spayed females.

Presenting signs
The presenting complaints were non-specific. All dogs were reported to be lethargic or depressed. Additional neurological signs were reported in two dogs; one was circling and head-pressing and one had a history of seizures. Three dogs had shown gastrointestinal signs but these were mild and intermittent and not considered sufficient to account for the electrolyte changes. Four of the dogs had received intravenous crystalloids on one or more occasions prior to referral and these had resulted in a temporary improvement in clinical signs.
All owners reported that their dog had a reduced or absent thirst. Five owners discovered that the dog would drink milk or fruit juice, but not water. One dog was reported to have deteriorated upon changing from a moist to a dry diet.
The only consistent abnormality on physical examination was the presence of dermatological changes. All dogs had hyperkeratosis, seborrhea and alopecia, particularly in the inguinal regions.

Urinalysis
The median urine specific gravity was 1.033 (range 1.018-1.046). There were no other significant findings on urinalysis.

Endocrine function
There were no consistent findings on endocrine function testing. Cortisol concentrations (pre-and post-adrenocorticotropin stimulation) were normal in 5/6 dogs, basal aldosterone concentrations were normal in 2/2 dogs, thyroid function was normal in 4/5 dogs, insulin-like growth factor 1 (IGF-1) was low in 3/3 dogs and one dog had a persistent hyperglycemia. The findings are summarized in Table 1.
One dog had reduced adrenal and thyroid function and was hyperglycemic.

Magnetic resonance imaging
Magnetic resonance imaging (MRI) was performed in four of the dogs and abnormalities were reported in two. Dog 4 had a marked hydrocephalus and in dog 5 the corpus callosum was absent. The MRI findings of dog 5 have been reported in more detail elsewhere (15). P. S. Chapman; D. Petrus; R. Neiger DOG/CAT Treatment Dog 4 was discharged from the hospital to receive oral fluids on an outpatient basis. The remaining seven dogs were admitted to the hospital and treated with intravenous crystalloid fluids. Three dogs were administered a balanced electrolyte solution containing 131 mmol/l sodium ([Na131]; Aqupharm No. 11 Compound Sodium Lactate Injection, Animalcare), two were administered 0.45% sodium chloride solution (equivalent to 74 mmol/l; [Na74]) and two were administered an isotonic solution of 0.18% sodium chloride (equivalent to 30 mmol/l sodium; [Na30]) and 4% glucose. The three dogs with the highest plasma sodium concentrations upon presentation (dogs 1, 2 and 8) were all administered the [Na131] fluid.
Amongst the seven dogs receiving intravenous fluids, the plasma sodium concentration decreased between 4 and 37 mmol/l during the first 24 hours of treatment. This corresponds to a rate of reduction in the plasma sodium of between 0.17 and 1.54 mmol/ l/hr (see Fig. 1). All three dogs administered the [Na131] fluid had a reduction in their plasma sodium of less than 0.5 mmol/l/hr. One dog each receiving fluids with lower sodium concentrations (dog 4, which received the [Na30] fluid and dog 5 which received [Na74] fluid) had reductions in their plasma sodium concentration of greater than 0.5 mmol/l/hr. Intravenous fluids were continued for between 2 and 15 days (median 7 days). The four dogs receiving the [Na30] and [Na74] fluids all had normalization of their serum sodium concentrations within 2 days. The dogs receiving the [Na131] fluid had a normalization of their sodium concentration at 4, 4 and 6 days of treatment. The progress in plasma sodium concentrations and the relationship to fluid type are shown in Figure 2.  Two dogs suffered complications whilst hospitalized. Dog 4 deteriorated neurologically, becoming profoundly depressed, during the first 24 hours of treatment. This dog had the greatest decrease in its sodium concentration, from 171 to 134 mmol/l during the first 24 hours of treatment (1.54 mmol/l/hr). Prior to admission to the hospital, the plasma sodium concentration had been documented to be 201 mmol/l. During the first day of treatment, the rate of fluid administration was reduced from 5.5 ml/kg/hr to 2.2 ml/ kg/hr and the fluid was changed from [Na30] to [Na74]. The sodium increased again to 169 before decreasing to 152 mmol/l over 4 days. The dog's neurological status improved but did not normalize and the dog remained markedly depressed. The owners elected to have the dog euthanased on the sixth day after admission. Dog 2 developed pneumonia whilst hospitalized and was euthanased on the seventh day after admission. Consent for a post mortem examination was not obtained in either of the euthanased dogs.
The five surviving dogs that were initially treated with intravenous fluids were gradually weaned onto their oral maintenance fluid requirement, which they accepted when it was mixed with food or milk. Dog 3, which was treated on an outpatient basis, also readily accepted oral fluids when they were mixed with food.
Two dogs were euthanased due to persistence of their original signs at 6 months and 15 months after diagnosis. One dog died from unrelated causes (motor vehicle trauma) 33 months after diagnosis. The three remaining dogs were stable at the time of their last followup at 6 months, 17 months and 17 months after diagnosis, although all owners reported continued abnormal mentation and behavior.

Discussion
Previous reports of primary adipsia in dogs have been of individual cases and this study represents the largest series of dogs reported with primary adipsia. Four of the nine cases previously reported in the literature have been Miniature Schnauzers (3,14,24,25). The other dogs described were a mixed breed dog (7), a Doberman (17), a Dalmatian (1) a Great Dane (13) and a Cairn Terrier (15). Interestingly none of the dogs diagnosed with primary adipsia in this study were Miniature Schnauzers. However, four of the eight dogs were Staffordshire bull terriers compared to only 2.1% of all dogs presenting to the Medicine Service at the QMHA between April 1999 and December 2002 (unpublished data).
Severe hypernatremia to the degree that these dogs exhibited is uncommon. However, before attributing the hypernatremia to adipsia, it is necessary to rule out other underlying causes. Hypernatremia can be caused by loss of hypotonic fluids, gain of solutes or a pure water deficit (5). The first two explanations seem unlikely in these patients. Hypotonic losses are most commonly from the gastrointestinal or urinary systems (5). Although three of the dogs had gastrointestinal signs, these were not considered of sufficient severity to cause electrolyte derangements of such a magnitude. Moreover, in a patient with mild gastrointestinal signs and normal access to water, hyponatremia would be a more likely sequela (5,22). Solutes may be gained through exogenous administration or endogenous retention, such as occurs with hyperaldosteronism. None of the dogs had any history of electrolyte administration and hyperaldosteronism is both rare in dogs and usually associated with significant hypokalemia (6). Thus, a pure water deficit is the most likely explanation for the observed hypernatremia. Pure water deficits may be caused by increased losses although it is uncommon for significant hypernatremia to develop in the presence of an intact thirst response and free access to water (5). The markedly increased fluid losses in diabetes insipidus may lead to hypernatremia if not accompanied by adequate compensatory polydipsia. However, not only did the dogs in this study not exhibit polyuria or polydipsia, the hallmark signs of diabetes insipidus, but the owners were consistent in their observations of a reduced or absent thirst. It has been proposed that lack of thirst in the presence of hypertonic dehydration is pathognomonic for primary adipsia (20).
The thirst response is mediated by a putative thirst centre in the hypothalamus (10). The most important stimulus for thirst is increased plasma osmolality and the cells in the thirst centre are assumed to act as osmoreceptors in a similar manner to, although functionally and anatomically separate from, those that stimulate ADH secretion. In addition, there are other non-osmotic stimuli for thirst such as hypotension, angiotensin II and dryness of the oropharyngeal mucous membranes (10). The effector limb of the thirst reflex also requires the integration of higher centres to coordinate the behavioural response. Thus, a defect in the thirst response might result from abnormal function of the osmoreceptors, a structural lesion of the hypothalamus or as a component of generalized central nervous system disease. Previously reported causes of adipsic hypernatremia in dogs include defective osmoregulation (3,7), hypothalamic dysplasia (1) and granulomatous meningoencephalitis (17).
It is likely that the dogs in this study represent a heterogeneous population without a single common etiology. Dog 4 was the only patient in which there was evidence of generalized central nervous system disease. The abnormal thirst response in this dog was most likely due to abnormal hypothalamic development secondary to congenital hydrocephalus. Adipsia secondary to hydrocephalus has been reported in a dog (7), a human (9) and a cat (8).
Gross evidence of generalized central nervous system disease was lacking in the other patients where MRI or cerebrospinal fluid analysis were performed, but cannot be excluded in the patients where these evaluations were not performed. Several owners did report persistent abnormalities in their pet's behavior and mentation, although they perceived the quality of life to be good. These persistent changes could be due to generalized central nervous system (CNS) disease or to the chronic effects of the lesions induced by hypernatremia, which may include parenchymal and subdural hemorrhages (19).
The Dalmatian dog in this study (dog 2) showed abnormalities in other endocrine functions that might be associated with a more complicated defect in hypothalamic function. A previous report has described adipsia in a Dalmatian dog which was attributed to a congenital malformation of the midline CNS structures (1), although this case did not exhibit any abnormalities of other endocrine functions. Imaging studies of the CNS were not performed in dog 2 and it is possible that a structural lesion may also have been present in our patient. The dog had low thyroxine levels, IGF-1 levels, an obtunded cortisol response to exogenous adrenocorticotropin and was persistently hyperglycemic, requiring exogenous insulin administration to establish and maintain normoglycemia. This latter finding is difficult to explain through abnormal hypothalamic or pituitary function alone.
In the remaining cases the most likely explanation for the adipsia was considered to be a localized defect in the osmoreceptors of the hypothalamic thirst centre. Moreover, it seems likely that the osmoreceptors controlling vasopressin secretion were also affected by a functional defect in some cases. Urine samples in all of the dogs demonstrated submaximally concentrated urine at the time of admission. This is an inappropriate response in a severely hypernatremic patient and suggests that either the osmoreceptors, ADH secretion or the response to ADH were abnormal. In the absence of ADH, production of hyposthenuric urine would be expected. However, this was not observed and in most cases the urine was moderately concentrated. This could be explained by hypovolemia-induced ADH secretion in the absence of osmoreceptor-induced secretion or by renal autoregulatory mechanisms. Thus, in addition to adipsia, these patients might have had a form of secondary central diabetes insipidus, attributable to defective osmotic-stimulated release of ADH. Unfortunately, ADH concentrations in response to volume and osmotic stimuli (7) were not measured to confirm this hypothesis.
Excepting the electrolyte changes, the findings on routine biochemistry were non-specific. Hypercholesterolemia was common and has been described previously in dogs with adipsic hypernatremia (7,13,24). The most remarkable finding on hematology was the presence of macrocytosis, which was also seen in one of the previously described dogs with adipsic hypernatremia (24). Macrocytosis has been described as an artefactual finding, associated with in vitro swelling of the hyperosmolar erythrocytes (4) and the authors believe that this was the cause of the macrocytosis in the cases described in this study.
In common with the cases presented here, those in the literature most often exhibited non-specific neurological signs (3,13,17,24,25). Seizures have been reported (7,25), although they affected only one dog in this series. Interestingly, the most consistent clinical finding was the presence of dermatological changes, which have been described previously (13,14). The etiology of these changes is unknown, but might be related to chronic dehydration of the stratum corneum. Progression of sodium concentrations over the period of hospitalization. Each line represents a case. Each case is identified by the type of fluid administered over the initial 24 hour period.
Rapid correction of hypernatremia will increase morbidity and mortality. In chronic hyperosmolar states, the central nervous system adapts by producing idiogenic osmoles such as myo-inositol (16). These allow the intracellular environment of the neurons to remain isotonic with respect to the extracellular fluid. However, correcting the hypernatremia faster than the idiogenic osmoles can be cleared may result in cellular swelling and neurological deterioration. It is therefore recommended that sodium concentration should be normalized over 48 hours (7) and that the rate of decrease of the sodium concentration not exceed 0.5 mmol/hr (23). This rate was exceeded in two patients in this study, one of which (dog 4) developed worsening neurological signs.
Although overzealous fluid administration was undoubtedly a factor in dog 4, it is possible that features of its disease may have predisposed it to a more rapid fall in its sodium concentration. In humans, two groups of patients with adipsia are recognized, with either partial or complete loss of the osmoreceptor function (20). If there is complete loss of osmoreceptor function, tonic ADH secretion cannot be suppressed when plasma osmolality falls, leading to continued and inappropriate ADH secretion (20) and an inability to excrete a free water load. This mechanism may also contribute to the development of hyponatremia in some affected dogs, emphasizing the need for judicious fluid therapy and close monitoring.
In the medium to long term, most dogs were able to function as acceptable pets, although continued abnormal mentation was reported. These persistent changes could be attributed to damage induced by the previous hypernatremia or by its treatment, neurological changes resulted from an underlying lesion or persistent hypernatremia.

Consequences for the clinics
Dogs with marked hypernatremia may have defective central hypothalamic osmoreceptors which cause a primary hypodypsia or adypsia, and owners need to be questioned carefully about the dogs drinking behaviour. Besides diagnostic imaging (magnetic resonance tomography) of the central nervous system to evaluate for potential structural abnormalities it would be ideal to measure plasma ADH levels during marked hypernatremia to assess for possible abnormal ADH secretion in response to osmotic stimuli. Increased sodium levels needs to be managed carefully with intravenous crystalloid fluids aiming for a drop in sodium of no more than 0.5 mmol/l/hr. Once sodium levels are normalised outpatient management may be achieved with increased water content in the diet with fair prognosis if no structural central nervous system abnormalities are present.