Glucocorticoid deficiency and syndrome of inappropriate antidiuresis: an underdiagnosed association?

Hyponatraemia is the commonest electrolyte disorder, which manifests in up to 20% of hospital inpatients. ¹ Approximately 40% of patients with hyponatraemia have the syndrome of inappropriate antidiuresis (SIAD), ² and there are a number of carefully considered guidelines for the assessment and management of patients with SIAD.^1,3–5 The original definition of SIAD contains the criterion that glucocorticoid deficiency should be excluded before the diagnosis is made, ⁶ and this criterion is reiterated in all of the recently published clinical guidelines.

Pure glucocorticoid deficiency leads to failure to effect renal free water clearance, with the development of water retention and dilutional hyponatraemia. Although plasma volume is expanded slightly, the patients present as euvolaemic to clinical examination; the biochemical picture is identical to that of SIAD, with hyponatraemia, normokalaemia and elevated urine osmolality and urine sodium concentration. This is in contrast to the clinical and biochemical characteristics of primary adrenal failure, where the combination of glucocorticoid and mineralocorticoid deficiency typically causes hypovolaemic hyponatraemia, which is accompanied by hyperkalemia. Although cortisol is necessary to facilitate water excretion from the kidney, the permissive effect on water excretion is only part of the pathogenesis of glucocorticoid deficiency. Patients who are glucocorticoid deficient also have elevated plasma concentrations of the antidiuretic hormone, vasopressin (AVP). ⁷ It is likely that the elevated plasma AVP concentrations are more important to the development of hyponatraemia in glucocorticoid deficiency; in adrenalectomized, but mineralocorticoid-replete rats, urinary dilution is almost completely restored by the administration of vasopressin receptor antagonists, ⁸ arguing strongly for a causal role for AVP.

However, published data strongly suggest that testing for glucocorticoid deficiency is not performed as frequently as recommended, and that this leads to the failure to identify steroid deficiency as a treatable cause of euvolaemic hyponatraemia. A retrospective review of routine clinical practice in 139 patients with all-cause hyponatraemia, who were admitted to a London teaching hospital, showed that only 33% of patients had assessment of cortisol secretion, with similar low levels of measurement of urine osmolality or sodium concentration. ⁹ In addition, the report of the Hyponatraemia Registry, an international observational study conducted in over 200 centres in the US and Europe, revealed that in 1524 patients specifically entered to the registry with a diagnosis of SIAD, only 33% had measurement of plasma cortisol concentrations. ¹⁰ Even within the context of a prospective single-site observational study of almost 500 patients with carefully defined SIAD, only 88% had appropriate tests to exclude glucocorticoid deficiency. ¹¹ Given that these studies were all designed and run in centres with a specific interest in hyponatraemia, it would not be unreasonable to assume that the rates of cortisol measurement are likely to be even lower in centres without a specific interest in hyponatraemia.

Although there are numerous case reports of hyponatraemia associated with hypocortisolaemia, the low rate of testing for plasma cortisol in published studies has dictated that the incidence of glucocorticoid deficiency in SIAD has until recently, remained undefined. A small retrospective study of 33 patients presenting to an endocrine unit with hyponatraemia showed that 40% of patients over the age of 65 years had steroid deficiency. ¹² A larger retrospective study of patients with hyponatremia admitted to an endocrine unit showed that 20% had hypopituitarism; in almost 90% of these patients, it was hyponatraemia which led to the diagnosis of hypopituitarism. ¹³ In the latter study, patients had been admitted up to four times previously to non-endocrine units before the diagnosis of hypopituitarism was made, and appropriate treatment with hydrocortisone was offered, which emphasizes the hazards of delayed diagnosis. Similar results were reported in a recent retrospective study from a specialized pituitary unit. Ten per cent of 260 patients admitted to a specialized tertiary referral pituitary unit were found to be hyponatraemic; hypopituitarism was found to be the cause of hyponatraemia in all cases. ¹⁴ Older patients with hypopituitarism were particularly likely to present with hyponatraemia. These studies were inherently biased by the small numbers and the referral to specialized endocrine units. Results from a recent prospective study have emphasized, however, that the diagnosis of hypopituitarism is not confined to patients presenting with SIAD to specialized endocrine units. Data from our own unit, reported by Cuesta et al., showed that 4% of a cohort of 500 unselected patients presenting to hospital with euvolaemic hyponatraemia had underlying steroid deficiency. ¹¹ In contrast to the patients presenting to specialized endocrine departments, only half of the patients had undiagnosed hypopituitarism. The other half had cortisol deficiency due to the effects of immunosuppressive steroid therapy. Some were admitted with hyponatraemia following discontinuation of oral steroids, whereas others were on maintenance steroids which had not been supplemented by stress dose glucocorticoids, to cope with the demands of acute illness. A small number were on inhaled steroids, and it is noteworthy that recent data have shown that 20% of patients on inhaled steroids fail the short synacthen test. ¹⁵ The lesson from these data is that failure to respond with rescue steroids during acute illness in patients on regular immunosuppressive steroids is just as likely to be the cause of presentation with hyponatraemia as hypopituitarism. This is particularly important, as patients on maintenance steroids for respiratory, rheumatology or gastroenterology conditions are highly unlikely to be aware of ‘sick day rules’, or to wear warning jewellery or carry steroid warning cards. ¹⁶

In neurosurgical units, acute ACTH/cortisol deficiency is now well recognized to be a clinically important cause of euvolaemic hyponatraemia, which presents with an identical biochemical pattern to SIAD. Half of all patients with subarachnoid haemorrhage develop hyponatraemia, ¹⁷ 90% have biochemistry typical of SIAD. Prospective studies, with daily measurement of plasma cortisol concentrations, have revealed that in 10% of ‘SIAD’ cases, hyponatraemia is due to acute pituitary injury with ACTH/cortisol deficiency. ¹⁸ A smaller proportion of patients with traumatic brain injury have hyponatraemia, at 15–20%, ¹⁹ but again, prospective studies have shown that up to 80% of these patients have transient hypocortisolaemia after brain injury, ²⁰ which may present with severe, symptomatic hyponatraemia. ²¹ Detection of cortisol deficiency offers the prospect of disease-specific treatment, as hydrocortisone therapy reverses hyponatraemia and improves patient wellbeing.

The ideal screening test for glucocorticoid deficiency remains debatable. Dynamic pituitary stimulation with insulin hypoglycaemia or glucagon testing is impractical, expensive and time consuming. We have used a cut-off of 300 nmol/L (as measured on the Beckman Coulter Unicell DXI 800 using the manufacturer’s reagents [Beckman Coulter, Brea, CA, USA]) on a random cortisol as a screening test in neurosurgical patients, in whom the acute nature of the hypopituitarism renders synacthen testing inaccurate ¹⁸ ; the cut-off was chosen to reflect the minimum stress levels of plasma cortisol concentration in patients recovering from vascular surgery. In patients who are stressed following neurosurgical intervention, we regard plasma cortisol concentrations of less than 300 nmol/L as clinically inappropriate, and treat empirically with hydrocortisone, with formal outpatient assessment of the hypothalamic-pituitary-adrenal (HPA) axis deferred until their clinical condition allows. ²² In patients admitted to non-neurosurgical units, plasma cortisol concentrations below 300 nmol/L are typically confirmed with synacthen stimulation, unless the patient has severe deficiency and is unwell, in which case immediate hydrocortisone therapy is commenced. In hyponatraemic patients whose random cortisol exceeds 300 nmol/L, synacthen testing has not revealed evidence of steroid deficiency. ¹¹ However, formal prospective validation of this approach is needed.

The challenge for clinicians is to identify those patients with hyponatraemia who have underlying glucocorticoid deficiency. As exclusion of cortisol deficiency is a necessary prerequisite to the diagnosis of SIAD, local protocols should emphasize that this is essential for the investigation of euvolaemic hyponatraemia. In routine practice, the need for a high index of suspicion has been recommended, in order not to miss hypopituitarism as a cause of hyponatraemia, ²³ in neurosurgical patients, associated hypoglycaemia or hypotension have been shown to be useful triggers to arouse suspicion of hypopituitarism as the cause of an SIAD-like presentation. ²⁴ The need to extend close consideration of the possibility of HPA axis suppression in patients taking immunosuppressive steroids is also important. Patients on long-term steroids are vulnerable to the development of functional adrenal insufficiency during acute illness, and hyponatraemia in these patients may be a marker for poor prognosis. Treatment with glucocorticoids is a cheap, efficient, and potentially life-saving option in patients with biochemical SIAD due to steroid deficiency.

Published data show that clinicians are not heeding the advice of published guidelines and measuring plasma cortisol in patients with SIAD. Clinical chemists may therefore have a valuable supportive role in this clinical scenario. Perhaps, it is timely to suggest that laboratory reports of hyponatraemia, with an SIAD-like picture, should be issued with a recommendation to check plasma cortisol. Collaboration between endocrinologists and laboratory staff to emphasize adherence to guidelines for the approach to hyponatraemia may be valuable in increasing the diagnosis of hyponatraemia due to steroid deficiency.