Mild water restriction with or without urea for the longterm treatment of syndrome of inappropriate antidiuretic hormone secretion (SIADH): Can urine osmolality help the choice?

BACKGROUND: Treatment options for chronic SIADH include water restriction (WR) and urea. The usefulness of urine osmolality to guide the choice of the treatment option is not clearly defined. We hypothesized that urine osmolality can indicate whether treatment with mild water restriction alone could be successful. METHODS: Retrospective Review of clinical and biochemical (blood and urine) data of patients with chronic SIADH treated for at least one year with mild WR (1.5-2l/day) either with or without urea. RESULTS: Twenty nine patients were included. Nine patients were treated by mild WR. Mean serum sodium (SNa) and mean Uosm were 129±2mEq/l and 274±78mOsm/kgH2O respectively before WR, and increased to 138.5±3mEq/l and 505±87mOsm/kgH2O (P<0.001). Eight patients were treated with mild WR and 15g urea daily, the SNa and Uosm before treatment were 127.5±3mEq/l and 340±100mOsm/kgH2O respectively and increased to 136.5±1mEq/l and 490±151mOsm/kgH2O (P<0.001). Four of the eight patients had a permanent low solute intake which contributed to hyponatremia. Twelve patients needed 30g urea daily combined with mild WR. The SNa and Uosm were respectively 126±2mEq/l and 595±176mOsm/kgH2O and increased to 136.5±2mEq/l and 698±157mOsm/kgH2O (P<0.05). Uosm increased in most of the treated patients. CONCLUSIONS: About 30% of patients could be treated by moderate WR alone. All these patients presented an initial urine osmolality lower than 400mOsm/kgH2O.

Low plasma bicarbonate level in hyponatremia related to adrenocorticotropin deficiency.

Patients with hyponatremia related to adrenocorticotropic deficiency are not easily distinguished by routine laboratory studies from patients with nonendocrine inappropriate secretion of antidiuretic hormone (SIADH). We wanted to investigate whether, in the routine biological analysis of such patients, some parameters could help to better identify this subgroup of hyponatremic patients. The biochemical profiles of 13 consecutive patients with hyponatremia related to ACTH deficiency were analyzed and compared with 30 consecutive patients with classical SIADH. Patients with adrenocorticotropic deficiency presented low uric acid and urea levels as in nonendocrine SIADH, but their total carbon dioxide was significantly lower (total CO(2), 20.5 +/- 3 vs. 25.5 +/- 2.4 mmol/liter; P < 0.001). Nine of the 13 patients presented a value lower than 22 mmol/liter, although this was not observed in the nonendocrine SIADH patients (P < 0.001). Arterial blood gas analysis was available in eight patients and showed a compensated respiratory alkalosis in most of them (pH 7.42 +/- 0.02; PCO(2), 30 +/- 5 mm Hg; HCO(3)(-), 20 +/- 2 mmol/liter; base excess, -3.4 +/- 1.8 mmol/liter). Aldosterone levels were much lower in ACTH deficiency patients during the hyponatremic state (33 +/- 40 pg/ml) when compared with the nonendocrine SIADH (120 +/- 60 pg/ml; P < 0.01). Correction of hyponatremia by cortisone therapy normalized total CO(2) and aldosterone levels. Low carbon dioxide level is a frequent observation in hyponatremia related to ACTH deficiency and could help to differentiate it from classical SIADH.

Evidence that chronicity of hyponatremia contributes to the high urate clearance observed in the syndrome of inappropriate antidiuretic hormone secretion.

The high fractional excretion (FE) of uric acid observed in hyponatremia associated with the syndrome of inappropriate secretion of antidiuretic hormone (SIADH) is commonly attributed to the volume-expanded state, although volume expansion in normonatremic volunteers is unable to increase urate clearance to a degree similar to that in SIADH. The goal of the present study is to analyze whether hyponatremia by itself could influence the FE of uric acid, as well as the effects of intravascular volume and glomerular filtration rate on FE of uric acid in SIADH. This study examines the effects of a 2-L infusion of isotonic saline over 24 hours on FE of uric acid in 9 normonatremic volunteers and 17 hyponatremic patients with SIADH. We also studied the FE of uric acid in 6 patients with SIADH with only mild water retention and the urate and creatinine clearances in 18 hyponatremic patients with SIADH before and after normalization of serum sodium levels by water restriction. When infusing 2 L of isotonic saline over 24 hours in healthy subjects, there was a decrease in plasma protein concentration of 8%, suggesting a similar degree of volume expansion than in patients with SIADH. The FE of uric acid did not increase to the same extent (9% +/- 1.5% versus 17% +/- 1.5%; P: < 0.01). Conversely, in 6 hyponatremic patients with mild water retention (1 L), the FE of uric acid was still high despite indirect signs of only a small increase in plasma volume. The mainstay of these observations is that chronicity of hyponatremia by itself could affect urate excretion. We also observed that in the patients with SIADH, high FE of uric acid inversely correlated with glomerular filtration rate (r = -0.66; P: < 0.01) only during the hyponatremic state. These data suggest that hyponatremia by itself, combined with mild volume expansion and glomerular filtration rate, has a role in the high FE of uric acid in the SIADH.

Azotemia (48 h) decreases the risk of brain damage in rats after correction of chronic hyponatremia.

Brain myelinolysis complicates excessive correction of chronic hyponatremia in man. Myelinolysis appear in rats for correction levels deltaSNa) > 20 mEq/l/24 h. We previously showed in rats that when chronic hyponatremia was corrected with urea, the incidence and the severity of brain lesions were significantly reduced compared to hypertonic saline. In man, hyponatremia is frequently associated with azotemia and hemo-dialysis usually corrects rapidly the serum sodium (SNa) but only few patients apparently develop demyelination. We hypothesize that uremic state protects brain against myelinolysis. This hypothesis was evaluated in rats developing azotemia by administration of mercuric chloride (HgCl2, 1.5 mg/kg). Severe (SNa < 120 mEq/l) hyponatremia (3 days) was induced by S.C. AVP and i.p. 2.5% D-glucose for 3 days. HgCl2 was injected on day 2. Hyponatremia was corrected on day 4 by i.p. injections of 5% NaCl in order to obtain a correction level largely above the toxic threshold for brain (deltaSNA approximately 30 mEq/l/24 h). Surviving rats were decapitated on day 10 for brain analysis. In the group with renal failure (Group I, n = 15, urea 59 mmol/l) the outcome was remarkably favourable with only three rats (3/15) dying before day 10 and only one of them (1/3) presenting myelinolysis-related neurologic symptoms. The 12 other rats (80%) survived in Group I without symptoms and brain analysis was normal in all of them despite large correction level (deltaSNa: 32 mEq/l/24 h). On the contrary in nine rats in which HgCl, did not produce significant azotemia (control 1, n = 9, urea: 11 mmol/l), all the rats developed severe neurologic symptoms and eight of them died before day 10. Similar catastrophic outcome was observed in the non-azotemic controls (control 2, no HgCl2 administration, n = 15, urea: 5 mmol/l). All of them developed myelinolysis-related neurologic symptoms and only four of them survived with severe brain lesions (survival 12/15 in Group I vs. 5/24 in pooled controls 1 and 2, p < 0.001). In conclusion, we showed for the first time that chronic hyponatremic rats with azotemia (48 h) tolerated large increases in SNa (approximately 30 mEq/l/24 h) without significant brain damage.

Therapeutic relowering of the serum sodium in a patient after excessive correction of hyponatremia.

BACKGROUND: Inappropriate correction of chronic hyponatremia could lead to major neuropathological sequelae. In man, the risk of brain myelinolysis increases strikingly when correction of the serum sodium exceeds 10-15 mEq/l/24 h. No treatment is actually available for this iatrogenic brain injury. However, recent experimental data showed that rapid reinduction of the hyponatremia greatly reduces the incidence of brain damage and death in case of serum sodium overshooting. SUBJECTS AND METHODS: We tested this rescue manoeuver in a 71-year-old woman with nausea, confusion and severe (SNa 106 mEq/l) chronic hyponatremia related to thiazides. It was associated with hypokalemia (SK: 3.2 mEq/l). RESULTS: Treatment with isotonic saline produced inappropriately high SNa correction level of +21 mEq/l after the first 24 h. After initial improvement, the neurological status deteriorated after 72 h. Rapid reinduction of the hyponatremia was then ordered. Administration of hypotonic fluids (by oral and i.v. route) combined with dDAVP induced a prompt decline in the SNa (-16 mEq/l/14 h) with a final gradient of correction of deltaSNa +9 mEq/l. This manoeuver was well tolerated without untoward effects. The natremia then progressively normalized and the patient completely recovered without neurological sequelae. CONCLUSION: Hypotonic fluids may be safely administered to decrease the natremia after excessive correction of hyponatremia for potential prevention of myelinolysis.

Restoration of the uricosuric effect of probenecid after triglycylvasopressine administration in a gouty patient.

A 35-year-old patient with severe gout and mild renal insufficiency presented very low urinary urate excretion. Volume expansion induced by fludrocortisone combined or not with a uricosuric drug (Benzbromarone) was unable to significantly increase his urate excretion. A combined Probenecid (PB) and Pyrazinamide (PZA) test was performed. These drugs being considered to affect renal tubular reabsorption or secretion. No significant modification of uric acid fractional excretion (FE.uric acid) was observed after PB and PZA. When the same test was performed after the administration of Triglycyl-lysine vasopressine (TGLV), a potent V1 receptor stimulator, we observed a three fold increase in FE.uric acid after PB intake (from 6 to 18%) followed by a decrease after PZA (from 18 to 5.6%). When TGLV was administered alone their was no significant modification of uric acid fractional excretion. We propose that TGLV decrease proximal tubular urate reabsorption that could only be detected when postsecretory reabsorption is blocked by an uricosuric drug.

Sarcoidosis of the paranasal sinuses treated with hydroxychloroquine.

A case of sarcoidosis of the paranasal sinuses is reported. Biopsies of the sinus mucosa showed typical noncaseating granulomas. Hydroxychloroquine, which is known to be active on the cutaneous form of sarcoidosis, was used here with success and is proposed as an effective alternative to high-dose systemic steroids.

Hyperuricemia as a clue for central diabetes insipidus (lack of V1 effect) in the differential diagnosis of polydipsia.

PURPOSE: In the differential diagnosis of patients with polyuria-polydipsia one must distinguish usually between primary polydipsia (PP) and central diabetes insipidus (CDI). The first situation is a state of volume expansion and the second of volume contraction. We evaluate whether serum uric acid determination could help to differentiate between the two conditions. PATIENTS AND METHODS: We analyzed the score of 13 consecutive patients with CDI, 7 patients with PP, and 7 patients with nephrogenic diabetes insipidus (NDI). Serum uric acid concentration was available during normonatremia without treatment with 1-desamino-8-D-arginine vasopressin (dDAVP), during mild dehydration and during treatment with dDAVP. In 8 of these patients plasma renin activity (PRA), urate, urea and creatinine clearances were also available. These data were also obtained in the patients with NDI. In 1 patient with CDI, we studied the effect on urate clearance of dDAVP, which stimulates exclusively the V2 receptors, and of triglycyl-lysine-vasopressin (TGLV), a potent V1-receptor agonist. RESULTS: Normonatremic polydypsic patients with CDI presented an increase in uric acid concentration (7.1 +/- 2.2 mg/dL), whereas in the PP group the value was decreased (3 +/- 0.75 mg/dL; P <0.001). All the normonatremic PP presented a serum uric acid concentration lower than 5 mg/dL, whereas all the normonatremic CDI patients, exept 1, presented a value higher than 5 mg/dL. In both groups blood urea concentration was decreased as a consequence of high renal clearances. The hyperuricemia of CDI was related to low uric acid clearances. Patients with hypernatremia and NDI presented a lower increase in serum uric acid concentration than those with similar levels of hypernatremia and CDI (NDI: 5.7 +/- 0.8 mg/dL and CDI: 7.9 +/- 2.3 mg/dL; P <0.05) and the NDI patients presented an urate clearance corrected for creatinine clearance which was significantly higher than in CDI (9% +/- 3% and 4% +/- 1.1%; P <0.01). When the patients with CDI were treated with dDAVP and normalyzed their PRA (0.9 +/- 0.4 ng/mL/h) we observed still mild hyperuricemia compared to controls (5.5 +/- 1.4 mg/dL and 4.3 +/- 0.9 mg/dL; P <0.01) and a low fractional excretion of filtered uric acid (6.5% +/- 1.7% compared to 8.2% +/- 2% in controls; P <0.05). Acute administration of dDAVP, stimulating the V2 receptors, in one patient with CDI, had no effect on urate clerance, while TGLV, which stimulates the V1 receptor, increased urate clearance. CONCLUSION: The presence of an serum uric acid concentration higher than 5 mg/dL in polyuric polydipsic patients is highly suggestive of CDI. Even when these patients are treated with dDAVP many of them remain hyperuricemic, and this seems to be the consequence of a lack of V1 receptor stimulation.

Lack of major hypoxia and significant brain damage in rats despite dramatic hyponatremic encephalopathy.

Brain myelinolysis could complicate the excessive correction of chronic hyponatremia. Recently it was suggested that hypoxia rather than correction of hyponatremia would be responsible for myelinolysis. We analyzed the incidence and the severity of potentially associated hypoxia and its consequences on survival and on the development of brain damage in rats in which major hyponatremic encephalopathy had developed after either pure acute hyponatremia (serum sodium concentration: -40 mEq/L/3 hr, group I, n = 8) or acute hyponatremia (serum sodium concentration: -30 mEq/L/3 hr, group II, n = 12) superimposed on chronic hyponatremia of 3 days' duration (serum sodium concentration: 113 mEq/L). Our study revealed the following: (1) Despite dramatic hyponatremic encephalopathy (convulsions, coma), hypoxia (PO2 < 70 mm Hg) was present, but the PO2 was not decreased below 40 mm Hg. All of these rats died rapidly if they remained hyponatremic. (2) In the animals rescued by NaCl, the incidence of brain myelinolysis was low (10%), whatever the duration (pure acute or chronic plus acute) of the hyponatremia and despite the combination of hypoxia with major hyponatremic encephalopathy. (3) When acute hyponatremia is superimposed on a chronic preexisting hyponatremic state, the acute component of serum sodium concentration decrease could be rapidly corrected (serum sodium concentration: +35 mEq/L/21 hr) without fear of permanent brain damage. Our results suggest that even in the presence of dramatic hyponatremic encephalopathy and associated hypoxia, neuropathologic sequelae are uncommon. Brain lesions related to post-anoxic encephalopathy probably develop only after respiratory arrest occurs.

In human patients, vascular water retention during DDAVP-related hyponatremia occurs mainly in the plasma volume and not in the erythrocyte.

DDAVP-related hyponatremia induces a blood volume expansion, but the analysis of fluid distribution in the vascular compartment has given controversial results in previous animal and human studies. In 5 healthy males, hyponatremia was induced by DDAVP and a free water intake during 3 days. Serum sodium concentration decreased from 138 +/- 0.8 mEq/L to 123 +/- 2.7 mEq/L on day 3. The plasma volume measured by dilution of marked albumin rose from 3033 +/- 230 ml to 3320 +/- 295 ml (p < 0.01). The mean corpuscular volume measured by microhematocrit increased slightly from 91.5 +/- 3.8 pl to 92.6 +/- 3.7 pl (p < 0.02). The red blood cell volume calculated with hematocrit and plasma volume did not change significantly (2565 ml to 2567 ml; not significant). In the present work, we demonstrated that in males the expansion of the plasma compartment almost completely amounted for the water retention in the intravascular volume. The erythrocyte volume increased only slightly, a finding that is consistent with an almost perfect adaptation of the erythrocyte cells to the hypoosmolality.

Therapeutic recommendations for management of severe hyponatremia: current concepts on pathogenesis and prevention of neurologic complications.

Patients with hyponatremia are exposed to major neurological complications. On the one hand hyponatremia itself produces brain edema, increased intracranial pressure which potentially leads to subsequent neuropathological sequelae or death. On the other hand excessive correction could be followed by development of brain demyelinating lesions (central pontine or extrapontine myelinolysis) with major disability or fatal outcome. Understanding of brain adaptative mechanisms to changes in osmolality largely contributes to explain these neurological events. When serum sodium decreases, the brain prevents swelling by extruding electrolytes and organic osmolytes, a process almost fully achieved after 48 h. Conversely, during subsequent increase in serum sodium, reestablishment of intracerebral osmolytes occurs but their reuptake is more delayed (+/- 5 days). In both circumstances, these mechanisms can be overwhelmed, leading to brain damage. Acute hyponatremia (< 48 h) is generally hospital-acquired, mainly in the postoperative state and/or after excessive fluid administration. After abrupt fall in serum sodium, seizure, respiratory arrest and coma may develop and these manifestations are sometimes explosive in nature. Recognition of even minor symptoms is crucial and implies prompt correction. There is generally no risk of brain myelinolysis in acute hyponatremia. Some factors are suspected to aggravate the prognosis of hyponatremic encephalopathy, including female gender (menstruant women), hypoxia and young age. Chronic hyponatremia (> 48 h) usually develops outside the hospital and is generally better tolerated. The risks of brain myelinolysis can be largely reduced by limiting the correction level to < or = 15 mEq/1/24 h. However, if necessary, the initial rate of correction can be rapid provided that the final correction remains < 15 mEq/1/24 h. However, when other recognized risk factors for myelinolysis (hypokalemia, liver disease, poor nutritional state, burns) are present, correction should not exceed 10 mEq/1/24 h. Demyelinization is also observed in hypernatremia but it follows greater (50%) increase in serum sodium than from hyponatremic baseline. For symptomatic hyponatremia, rapid correction is usually obtained by hypertonic saline (3%) infusion. Another option consists in administration of intravenous or oral urea. Urea allows a rapid reduction of brain edema and intracranial pressure which is followed by subsequent correction of hyponatremia. Experimental data also suggest that treatment of hyponatremia with urea is associated with a lower incidence of myelinolysis. In hyponatremic patients without symptoms, there is no need for rapid correction and the treatment should be more conservative. Close monitoring of the serum sodium is indicated initially and if necessary, correction must be stopped and diuresis interrupted with dDAVP. Given recent experimental data, in patients overly corrected (delta SNa > 15 mEq/1/24 h), the risk of myelinolysis could be greatly reduced by rapidly decreasing the serum sodium through hypotonic fluids administration and dDAVP.

Evidence in hyponatremia related to inappropriate secretion of ADH that V1 receptor stimulation contributes to the increase in renal uric acid clearance.

In hyponatremia related to syndrome of inappropriate antidiuretic hormone (SIADH), hypouricemia is explained primarily by the high uric acid clearance rate that results from the decrease in tubular uric acid reabsorption. This modification of tubular handling of uric acid is considered to be induced by the increase in the "effective vascular volume". This study was designed to determine if V1-receptor stimulation participates in the development of a high uric acid clearance rate as in SIADH, in which the antidiuretic hormone acts on V1 and V2 receptors. Therefore, the urate clearance rate was measured in seven volunteers with 1-desamino-8-D-arginine vasopressin (dDAVP)-induced hyponatremia, with dDVAP stimulating exclusively the V2 receptors (Group I), and in six patients with SIADH (Group II) during both normo- and hyponatremia. As expected, in both groups, the serum uric acid concentration decreased during hyponatremia, but did so to a larger extent in the patients with SIADH (-53% versus -29%, P < 0.02). Despite similar levels of hyponatremia (126 +/- 5 mmol/L and 125 +/- 5.5 mmol/L), of hypoproteinemia (64 +/- 5 g/L and 63 +/- 5 g/L) and of salt excretion (FENa, 0.66 +/- 0.28% and 0.73 +/- 0.25%), the urate clearance (8.3 +/- 3.3 mL/min) and the fractional excretion of filtered uric acid (5.7 +/- 2%) in Group I were not significantly different during hyponatremia than during normonatremia (6.4 +/- 1.5 mL/min and 5.4 +/- 0.9%). On the other hand, in Group II, both parameters were increased (17.8 +/- 2.9 mL/min and 19.6 +/- 5.3%; P < 0.001) and both values were higher than in the dDAVP-induced hyponatremia (P < 0.01). Additionally, the administration of a potent V1-receptor agonist (triglycyl-lysine-vasopressin) in a patient with central diabetes insipidus with preexisting dDAVP-induced hyponatremia produced a rapid increase of urate clearance. Because dDAVP acts only on the V2 receptors, these data suggest that the higher urate clearance observed during hyponatremia related to SIADH is not only the consequence of an increased "effective vascular volume," but that V1-receptor stimulation also contributes to it, by a mechanism that remains to be determined.

Reinduction of hyponatremia improves survival in rats with myelinolysis-related neurologic symptoms.

Brain myelinolysis occurs after excessive correction (delta SNa > 20 mEq/1/24 hours) of chronic hyponatremia. However, we showed recently that the mechanisms leading to brain myelinolysis remain reversible. Indeed, reinduction of the hyponatremia by water administration despite 12 hours of sustained excessive correction could prevent the development of demyelination in rats still asymptomatic at that time. Whether this therapeutic maneuver could be also beneficial to rats with preexisting myelinolysis-related neurologic symptoms is unknown. Therefore we evaluated here the effect of reinduction of the hyponatremia on the survival and on brain damage in rats presenting obvious neurologic symptoms after excessive correction of hyponatremia. After 3 days of severe hyponatremia induced by 2.5 D-glucose in water and continuous infusion of AVP, rats were submitted to a large correction (delta SNa approximately 30 mEq/l) by 2 i.p. injections of hypertonic saline given over 24 hours. In group I (n = 15) the rats developing neurologic symptoms during the first 24 hours of correction received one i.p. injection of distilled water which rapidly decreased the natremia to a final correction gradient <20 mEq/l/24 hour. In group II (n = 13, controls) the symptomatic rats were left permanently overcorrected. In group I, after water administration, the neurological manifestations were generally attenuated or disappeared. Seven of the 15 rats (47%) in this group survived up to day 10 with a mean survival time of 7.5 +/- 2 days, an outcome clearly improved as compared to group II (controls): only 1 of the 13 rats (7%, p < 0.03) was still alive on day 10 and the mean survival time was 3.3 +/- 2 days (p < 0.001) in this group II. The duration of the symptoms also influences the prognosis. In group I, in 9 rats the water administration was performed 4 hours after symptoms onset. These rats had a better outcome than the 6 rats with more sustained (8-10 hours) neurologic symptoms before water loading. Brain analysis in the 7 surviving rats of group I demonstrated demyelinating lesions in only 2 of them, suggesting the reversibility of the process even when neurologic manifestation developed. In conclusion, after exposure to an excessive correction of chronic hyponatremia, even when rats have developed myelinolysis-related neurologic symptoms, hypotonic fluids administration could improve survival and could prevent the subsequent development of brain myelinolysis.

Brain myelinolysis following hypernatremia in rats.

Brain myelinolysis could develop after excessive correction (delta SNa > 20-25 mEq/1/24 hour [h]) of chronic hyponatremia; however, this neurological event is not recognized as a complication of hypernatremia when arising from a normonatremic baseline. Previous animal studies were unable to reproduce these brain lesions in hypernatremia after acute increase of serum sodium to moderately hypernatremic levels. We hypothesize that to produce brain dehydration and myelinolysis from normonatremic baseline requires a more important osmotic gradient than when starting from hyponatremic state. Rapid and sustained hypernatremia (at least > 6 to 12 h) was induced in male rats by i.p. administration of NaCl 2 M (3 injections at 6 h intervals). The NaCl doses were determined to define two groups of hypernatremic rats (moderate and severe hypernatremia) for further analysis of the neurological outcome. In group 1 (moderate hypernatremia, n = 26) 8 rats died early (< 12 h) after the beginning of the NaCl administration without specific neurologic manifestations. All the surviving rats fared well and were asymptomatic at time of death (day 8). They were submitted for at least 6 to 12 h to a serum sodium gradient of 28 +/- 6 mEq/l. Brain analysis was normal in all of them without brain demyelinating lesions. In group 2 (n = 51), 24 rats also died rapidly (< 12 h). The surviving rats developed severe neurologic symptoms as typically encountered in hyponatremic rats with myelinolysis. The majority of them died before day 8. The hypernatremic gradient in this group was significantly higher than rats in group 1 that completely recovered (mean delta SNa: 39 +/- 8 mEq/l, p < 0.001). In the 7 surviving rats (mean delta SNa: 33 +/- 3 mEq/l) brain analysis demonstrated severe demyelinating lesions similar to the histologic changes observed in hyponatremia-related myelinolysis. We demonstrated for the first time that high and sustained levels of hypernatremia could induce brain myelinolysis and that the osmotic gradient necessary to produce brain lesions is higher for normonatremic than for hyponatremic rats.

Restoration by corticosteroids of the hyperaldosteronism in hyponatraemic rats with panhypopituitarism.

1. In the syndrome of inappropriate secretion of antidiuretic hormone, hyponatraemia is associated with a normal bicarbonate concentration despite dilution. This normal bicarbonate concentration is related to the development of a hyperaldosteronism, which is attributed to a direct stimulation of the zona glomerulosa by the hyponatraemic state. Some workers have suggested that, to develop this hyperaldosteronism requires the presence of a pituitary factor. To determine whether the pituitary gland plays a role in this hyponatraemia-induced hyperaldosteronism, water intoxication was performed for 24 h in normal and in panhypopituitaric rats. 2. In normal rats, hyponatraemia (108 mmol/l), induced by the administration of 1-desamino-8-D-arginine vasopressin and 2.5% D-glucose-0.45% NaCl by gavage (15% body weight) was associated with a mild increase in bicarbonate concentration, and blood acid-base equilibrium showed a mixed metabolic and respiratory alkalosis (pH 7.57, partial pressure of CO2 29 mmHg, base excess +5.5 mmol/l), and aldosterone concentration was increased 3-fold as compared with the control value. When hyponatraemia (110 mmol/l) was induced in a similar manner in panhypopituitaric rats, we observed a very low aldosterone concentration (< 50 pg/ml) and a compensated respiratory alkalosis (pH 7.45, partial pressure of CO2 30 mmHg, base excess -2.6 mmol/l). The restoration of a hyperaldosteronaemic state in this group of rats was related essentially to corticosteroid intake. 3. These data suggest that corticosteroids play a critical role in the development of hyponatraemia-related hyperaldosteronism, a phenomenon not necessarily dependent on a pituitary factor.