Tissue Glucocorticoid Metabolism in Adrenal Insufficiency: A Prospective Study of Dual-release Hydrocortisone Therapy

BACKGROUND

Patients with adrenal insufficiency (AI) require life-long glucocorticoid (GC) replacement therapy. Within tissues, cortisol (F) availability is under the control of the isozymes of 11β-hydroxysteroid dehydrogenase (11β-HSD). We hypothesize that corticosteroid metabolism is altered in patients with AI because of the nonphysiological pattern of current immediate release hydrocortisone (IR-HC) replacement therapy. The use of a once-daily dual-release hydrocortisone (DR-HC) preparation, (Plenadren®), offers a more physiological cortisol profile and may alter corticosteroid metabolism in vivo.

STUDY DESIGN AND METHODS

Prospective crossover study assessing the impact of 12 weeks of DR-HC on systemic GC metabolism (urinary steroid metabolome profiling), cortisol activation in the liver (cortisone acetate challenge test), and subcutaneous adipose tissue (microdialysis, biopsy for gene expression analysis) in 51 patients with AI (primary and secondary) in comparison to IR-HC treatment and age- and BMI-matched controls.

RESULTS

Patients with AI receiving IR-HC had a higher median 24-hour urinary excretion of cortisol compared with healthy controls (72.1 µg/24 hours [IQR 43.6-124.2] vs 51.9 µg/24 hours [35.5-72.3], P = .02), with lower global activity of 11β-HSD2 and higher 5-alpha reductase activity. Following the switch from IR-HC to DR-HC therapy, there was a significant reduction in urinary cortisol and total GC metabolite excretion, which was most significant in the evening. There was an increase in 11β-HSD2 activity. Hepatic 11β-HSD1 activity was not significantly altered after switching to DR-HC, but there was a significant reduction in the expression and activity of 11β-HSD1 in subcutaneous adipose tissue.

CONCLUSION

Using comprehensive in vivo techniques, we have demonstrated abnormalities in corticosteroid metabolism in patients with primary and secondary AI receiving IR-HC. This dysregulation of pre-receptor glucocorticoid metabolism results in enhanced glucocorticoid activation in adipose tissue, which was ameliorated by treatment with DR-HC.

The effects of acute hyponatraemia on bone turnover in patients with subarachnoid haemorrhage: A preliminary report

CONTEXT

Animal data and cross-sectional human studies have established that chronic hyponatraemia predisposes to osteoporosis; the effects of acute hyponatraemia on bone turnover have not been determined. Our objective was to test the hypothesis that acute hyponatraemia leads to dynamic effects on bone turnover.

DESIGN

A prospective observational pilot study.

METHODS

Bone turnover markers [C-terminal crosslinking telopeptide of type 1 collagen (CTX-1), N-propeptide of type 1 collagen (P1NP) and osteocalcin] were measured prospectively over one week in 22 eunatraemic patients with subarachnoid haemorrhage. Patients treated with glucocorticoids were excluded.

RESULTS

Eight patients developed acute hyponatraemia, median nadir plasma sodium concentration 131 mmol/L (IQR 128-132), and 14 remained eunatraemic, nadir plasma sodium concentration 136 mmol/L (IQR 133-137). Significant main effects of hyponatraemia were found for P1NP (p = .02) and P1NP:CTX-1 ratio (p = .02), both fell in patients with acute hyponatraemia, with significant interaction between hyponatraemia and time from baseline for P1NP (p = .02). Significant main effects of time from baseline (p < .001) but not hyponatraemia (p = .07) were found for osteocalcin. For CTX-1, significant main effects of time from baseline (p = .001) but not hyponatraemia (p = .65) were found. There was a positive correlation between change in P1NP:CTX-1 ratio and nadir plasma sodium concentration, r = +.43, p = .04. Median serum cortisol (measured on days 1, 3 and 7) was higher in the hyponatraemia group than in those who remained eunatraemic, 545 nmol/L (IQR 373-778) versus 444 nmol/L (IQR 379-542) p = .03.

CONCLUSION

These data suggest that acute mild hyponatraemia is associated with a reduction in bone formation activity.

Cardiometabolic and psychological effects of dual-release hydrocortisone: a cross-over study

BACKGROUND

Adrenal insufficiency (AI) is associated with increased cardiovascular morbidity and mortality and reduced quality of life (QoL). Optimum glucocorticoid (GC) dosing and timing are crucial in the treatment of AI, yet the natural circadian secretion of cortisol is difficult to mimic. The once-daily dual-release hydrocortisone (DR-HC) preparation (Plenadren®), offers a more physiological cortisol profile and may address unmet needs.

METHODS

An investigator-initiated, prospective, cross-over study in patients with AI. Following baseline assessment of cardiometabolic risk factors and QoL, patients switched from their usual hydrocortisone regimen to a once-daily dose equivalent of DR-HC and were reassessed after 12 weeks of treatment.

RESULTS

Fifty-one patients (21 PAI/30 SAI) completed the study. Mean age was 41.6 years (s.d. 13), and 58% (n = 30) were male. The median daily HC dose before study entry was 20 mg (IQR 15-20 mg). After 3 months on DR-HC, the mean SBP decreased by 5.7 mmHg, P = 0.0019 and DBP decreased by 4.5 mmHg, P = 0.0011. There was also a significant reduction in mean body weight (-1.23 kg, P = 0.006) and BMI (-0.3 kg/m2, P = 0.003). In a sub-analysis, there was a greater reduction in SBP observed in patients with SAI when compared to PAI post-DR-HC. Patients reported significant improvements in QoL using three validated QoL questionnaires, with a greater improvement in PAI.

CONCLUSION

Dual-release hydrocortisone decreases BP, weight and BMI compared with conventional HC treatment, even at physiological GC replacement doses. Additionally, DR-HC confers significant improvements in QoL compared to immediate-release HC, particularly in patients with PAI, which is also reflected in the patient preference for DR-HC.

Fluid Restriction Therapy for Chronic SIAD; Results of a Prospective Randomized Controlled Trial

CONTEXT

Fluid restriction (FR) is the recommended first-line treatment for syndrome of inappropriate antidiuresis (SIAD), despite the lack of prospective data to support its efficacy.

DESIGN

A prospective nonblinded randomized controlled trial of FR versus no treatment in chronic SIAD.

INTERVENTIONS AND OUTCOME

A total of 46 patients with chronic asymptomatic SIAD were randomized to either FR (1 liter/day) or no specific hyponatremia treatment (NoTx) for 1 month. The primary endpoints were change in plasma sodium concentration (pNa) at days 4 and 30.

RESULTS

Median baseline pNa was similar in the 2 groups [127 mmol/L (interquartile range [IQR] 126-129) FR and 128 mmol/L (IQR 126-129) NoTx, P = 0.36]. PNa rose by 3 mmol/L (IQR 2-4) after 3 days FR, compared with 1 mmol/L (IQR 0-3) NoTx, P = 0.005. There was minimal additional rise in pNa by day 30; median pNa increased from baseline by 4 mmol/L (IQR 2-6) in FR, compared with 1 mmol/L (IQR 0-1) NoTx, P = 0.04. After 3 days, 17% of FR had a rise in pNa of ≥5 mmol/L, compared with 4% NoTx, RR 4.0 (95% CI 0.66-25.69), P = 0.35. After 3 days, 61% of FR corrected pNa to ≥130 mmol/L, compared with 39% of NoTx, RR 1.56 (95% CI 0.87-2.94), P = 0.24.

CONCLUSION

FR induces a modest early rise in pNa in patients with chronic SIAD, with minimal additional rise thereafter, and it is well-tolerated. More than one-third of patients fail to reach a pNa ≥130 mmol/L after 3 days of FR, emphasizing the clinical need for additional therapies for SIAD in some patients.

SUN-352 The Effects of Acute Hyponatremia on Bone Remodeling Markers in Patients with Subarachnoid Hemorrhage

Animal data and cross-sectional human studies have established that chronic hyponatremia predisposes to osteoporosis; the effects of acute hyponatremia on bone remodeling are unknown. Serum markers of bone remodeling (total procollagen type 1 amino-terminal propeptide (P1NP), bone specific alkaline phosphatase (bone ALP), N-mid-osteocalcin (OCI) and C-terminal teleopeptides of type I collagen (CTX-1)) were assessed in a cohort of patients admitted with subarachnoid hemorrhage (SAH), who were prospectively studied over seven days. The ratio of P1NP:CTX-1 was calculated to report a bone formation index.

Twenty-two patients (13 women), median (IQR) age 53 (47, 62) years were recruited. Patients who developed post-SAH ACTH deficiency and those treated with glucocorticoids, or continuous enteral feeding were excluded. All patients were eunatremic on initial assessment. Eight patients developed acute hyponatremia, median nadir plasma sodium concentration (pNa) 131 (128, 132) mmol/L, and 14 remained eunatremic, nadir pNa 136 (133, 137) mmol/L. The groups were matched for age, 25-hydroxy Vitamin D, PTH, WFSS and Fischer scores. Serum cortisol concentration was greater in the hyponatremic group, 571 (504, 671) nmol/L, than the eunatremic group, 449 (400, 501) nmol/L, p=0.008. Bone remodeling markers and bone formation index (P1NP:CTX-1 ratio) were similar in the two groups at baseline.

There was a significant rise in CTX-1 in both hyponatremic patients, +0.15 (0.09, 0.37) μg/l, p = 0.009, and patients who remained eunatremic, +0.11 (-0.02, 0.23) μg/l, p = 0.04, with no significant difference between the groups. There was, however, a significant fall in P1NP:CTX-1 ratio in patients with acute hyponatremia, p = 0.02, but no significant change in eunatraemic patients, with significant between group difference, p = 0.02.

Changes in P1NP and OCI correlated positively with nadir pNa; r = 0.43, p = 0.04 and r = 0.61, p = 0.001 respectively. In addition, there was a positive correlation between change in P1NP:CTX-1 ratio and nadir pNa, r = 0.43, p = 0.04. There was no correlation between change in OCI or CTX-1 and nadir pNa. Serum cortisol was strongly negatively correlated with change in P1NP (r = -0.64, p = 0.001) but not with change in other bone remodeling markers.

Acute hyponatremia following SAH is associated with a fall in bone formation index; physiological hypercortisolemia may contribute to this. Further analysis with larger numbers will help us determine whether hyponatremia is an independent risk factor.

Continuous Versus Bolus Infusion of Hypertonic Saline in the Treatment of Symptomatic Hyponatremia Caused by SIAD

BACKGROUND

Acute hyponatremia is a medical emergency that confers high mortality, attributed primarily to cerebral edema. Expert guidelines advocate the use of intravenous boluses of hypertonic saline rather than traditional continuous infusion to achieve a faster initial rise in plasma sodium (pNa) concentration. However, there is a limited evidence base for this recommended policy change.

METHODS

We prospectively assessed the clinical and biochemical outcomes in patients treated for symptomatic hyponatremia caused by syndrome of inappropriate antidiuresis in response to intravenous bolus treatment with 3% saline (100 mL, repeated up to two more times) and compared the outcomes to retrospective data from patients treated with continuous intravenous infusion of low-dose (20 mL/h) 3% saline.

RESULTS

Twenty-two patients were treated with bolus infusion and 28 with continuous infusion. Three percent saline bolus caused more rapid elevation of pNa at 6 hours [median (range) 6 (2 to11) vs 3 (1 to 4) mmol/L, P < 0.0001], with a concomitant improvement in Glasgow Coma Scale (GCS) [median (range) 3 (1 to 6) vs 1 (-2 to 2), P < 0.0001] at 6 hours. Median pNa concentration was similar at 24 hours in the two treatment groups. The administration of a third saline bolus was associated with greater need for dextrose/dDAVP to prevent overcorrection (OR 24; P = 0.006). There were no cases of osmotic demyelination in either group.

CONCLUSION

Three percent saline bolus produces faster initial elevation of pNa than continuous infusion with quicker restoration of GCS, and without osmotic demyelination. Frequent electrolyte monitoring, and judicious intervention with dDAVP is required to prevent overcorrection with bolus therapy.

Octreotide use for rescue of vision in a pregnant patient with acromegaly

Pregnancy in acromegaly is rare and generally safe, but tumour expansion may occur. Managing tumour expansion during pregnancy is complex, due to the potential complications of surgery and side effects of anti-tumoural medication. A 32-year-old woman was diagnosed with acromegaly at 11-week gestation. She had a large macroadenoma invading the suprasellar cistern. She developed bitemporal hemianopia at 20-week gestation. She declined surgery and was commenced on 100 µg subcutaneous octreotide tds, with normalisation of her visual fields after 2 weeks of therapy. She had a further deterioration in her visual fields at 24-week gestation, which responded to an increase in subcutaneous octreotide to 150 µg tds. Her vision remained stable for the remainder of the pregnancy. She was diagnosed with gestational diabetes at 14/40 and was commenced on basal bolus insulin regimen at 22/40 gestation. She otherwise had no obstetric complications. Foetal growth continued along the 50th centile throughout pregnancy. She underwent an elective caesarean section at 34/40, foetal weight was 3.2 kg at birth with an APGAR score of 9. The neonate was examined by an experienced neonatologist and there were no congenital abnormalities identified. She opted not to breastfeed and she is menstruating regularly post-partum. She was commenced on octreotide LAR 40 mg and referred for surgery. At last follow-up, 2 years post-partum, the infant has been developing normally. In conclusion, our case describes a first presentation of acromegaly in pregnancy and rescue of visual field loss with somatostatin analogue therapy. Learning points: Tumour expansion may occur in acromegaly during pregnancy. Treatment options for tumour expansion in pregnancy include both medical and surgical options. Somatostatin analogues may be a viable medical alternative to surgery in patients with tumour expansion during pregnancy.

OR16-6 Cardiometabolic Abnormalities in Patients with Acromegaly with Elevated Plasma IGF-1 Concentrations but GH Concentrations <2ng/ml

The increased cardiovascular mortality in acromegaly can be reversed by achieving plasma GH concentrations to a target of <2ng/ml and normalizing IGF-1 concentrations to within the age-related reference range. However, in approximately 30% of patients, plasma GH concentrations which are on target are associated with discordantly elevated plasma IGF-1 concentrations. It is not known whether this discordance confers excess morbidity. The aim of this study was to determine whether discordance of plasma GH and IGF-1 negatively affects parameters which predict cardiovascular morbidity. Methods: Group 1 (Concordant): 28 patients (15F,13M) with plasma GH <2ng/ml and IGF-1 concentrations within the age and sex matched reference range. Group 2 (Discordant): 27 patients (5F, 22M) with plasma GH <2ng/ml and elevated plasma IGF-1 concentrations above the age and sex-matched reference range. IGF-1 and GH concentrations were analysed in one laboratory using IDS-iSYS assay. Patients were excluded if they were receiving Pegvisomant therapy, if they had a change in therapy in the preceding 6 months or if the random GH was >2ng/ml. Parameters Measured: OGTT with GH, fasting lipid profile, HOMA-IR, Carotid Intima Thickness (CIMT), pulse wave velocity (PWV), flow mediated dilatation (FMD) and 24-hour ambulatory blood pressure. All tests were completed with a standardised protocol over one day. IGF-1 concentrations were calculated as % Upper limit of normal (%ULN) to allow comparison across different age matched reference ranges. Results: The median age (range) was similar in the two groups; 54 (32-78) years in the concordant group and 61 (34-83) years in the discordant group, p=0.45. There were no significant between group differences in BMI, smoking, hypertension, diabetes mellitus or obstructive sleep apnoea. GH nadir during OGTT (p=0.03) and basal plasma GH (p=0.03) were higher in the discordant group. HOMA-IR was also higher in the discordant group (median = 1.75, range 0.73 - 4.88) compared to the concordant group (median = 1.123, range 0.27-4.63), p=0.01. However, there were no significant differences in HbA1c and fasting glucose between the two groups. Furthermore, there were no significant differences between IMT (p=0.63), PWV (p=0.3), mean daytime (p=0.48) or night-time systolic blood pressure (p=0.54) between the two groups. Conclusion: Elevated plasma IGF-1, with controlled GH concentrations, was associated with higher basal GH, higher nadir GH on OGTT and higher markers of insulin resistance in comparison to patients with controlled GH and IGF-1 concentrations. However, we found no difference in other predictors of future cardiovascular risk between the two groups. Further investigation of these findings is warranted to determine if the differences in HOMA-IR and GH dynamics between the two groups requires intervention.

Pregnancy in acromegaly is safe and is associated with improvements in IGF-1 concentrations

Pregnancy is rarely reported in acromegaly. Many patients are diagnosed in later life and younger patients may have subfertility due to hypopituitarism. We present a case series of 17 pregnancies in 12 women with acromegaly. Twelve women with acromegaly who completed pregnancy were identified from centres involved in the Irish Pituitary Study. Eleven women had pituitary macroadenomas and one woman had a microadenoma. Only 5/17 pregnancies had optimal biochemical control of acromegaly preconception, as defined by IGF-1 concentration in the age-related reference level and plasma GH concentration of <2 μg/L. In 6/17 pregnancies, dopamine agonist treatment was continued during pregnancy; all other acromegaly treatments were discontinued during pregnancy. Effect of pregnancy on acromegaly: No patient developed new visual field abnormalities, or symptoms suggestive of tumour expansion during pregnancy. In 9/12 patients, plasma IGF-1 concentrations that were elevated preconception normalised during pregnancy. There was a reduction in plasma IGF-1 concentrations, though not into the normal range, in a further two pregnancies. Effect of acromegaly on pregnancy: 15 healthy babies were born at term; one patient underwent emergency C-section at 32 weeks for pre-eclampsia, and one twin pregnancy had an elective C-section at 35 weeks' gestation. Blood pressure remained within normal limits in the remainder of the pregnancies. Gestational diabetes did not develop in any pregnancy. Our data suggests that pregnancy in women with acromegaly is generally safe, from a maternal and foetal perspective. Furthermore, biochemical control tends to improve despite the withdrawal of somatostatin analogue therapy during pregnancy.

Clinical features and autoimmune associations in patients presenting with Idiopathic Isolated ACTH deficiency

OBJECTIVE

Idiopathic Isolated ATCH deficiency (IIAD) is a rare cause of secondary adrenal insufficiency. As the condition is rare, and the diagnostic criteria ill-defined, there are few good clinical descriptions in the literature. We have described presenting features, autoimmune associations, natural history and responses to CRF, in a large case series of patients presenting with IIAD.

DESIGN

This is a retrospective case note analysis with data derived from the recently commenced National Pituitary Database of Ireland.

PATIENTS

Twenty-three patients with isolated ACTH deficiency were identified. A thorough chart and biochemistry review was performed.

RESULTS

Twenty-three patients were examined (18 women and 5 men). Age at presentation ranged from 17 to 88 years, (median 48 years). Most patients complained of fatigue; 9 patients presented with hyponatraemia, 13 had autoimmune illnesses (primary hypothyroidism, n = 9). CRF stimulation testing was available in 12 of the 23 patients, 5 of whom demonstrated a rise in plasma ACTH concentrations, indicating hypothalamic, rather than pituitary aetiology. Two patients recovered ACTH secretion, and 2 patients progressed to have other pituitary hormone deficiencies.

CONCLUSIONS

IIAD typically presents with insidious symptoms. Euvolaemic hyponatraemia is common at diagnosis. It is associated with autoimmune diseases, particularly primary hypothyroidism. As two patients recovered ACTH secretion, and two progressed to other pituitary hormone deficits, repeat pituitary testing should be considered, to identify recovery of function, or progression to other hormone deficits.

IgG4 hypophysitis - a rare and underdiagnosed cause of pituitary gland and stalk mass-like thickening

Our aim is to present a typical case of IgG4-related hypophysitis, which will offer insight into the aetiology and pathogenesis of this relatively newly described disease. IgG4 Related Disease is a protean systemic condition that mimics inflammatory, infectious, and malignant processes. Biopsy of affected organs will show a typical histopathological pattern.

Mortality rates are lower in SIAD, than in hypervolaemic or hypovolaemic hyponatraemia: Results of a prospective observational study

OBJECTIVE

Hyponatraemia is associated with increased mortality, but the mortality associated specifically with SIAD is not known. We hypothesized that mortality in SIAD was elevated, but that it was less than in hypervolaemic (HEN) or hypovolaemic (HON) hyponatraemia.

DESIGN

Mortality rates are presented as risk ratios (RR),with 95% confidence intervals (CI), and compared to normonatraemic controls (NN).

METHODS

Prospective, single centre, noninterventional study of all patients with hyponatraemia (≤130 mmol/L) admitted to hospital.

RESULTS

A total of 1323 admissions with hyponatraemia were prospectively evaluated and 1136 contemporaneous NN controls. 431(32.6%) hyponatraemic patients had HON, 573(43.3%) had SIAD and 275(20.8%) patients had HEN. In patient mortality was higher in hyponatraemia than NN (9.1% vs 3.3%, P<.0001). The RRs for in-hospital mortality compared to NN were: SIAD, 1.76 (95% CI 1.08-2.8, P=.02), HON 2.77 (95% CI 1.8-4.3, P<.0001) and HEN, 4.9 (95% CI 3.2-7.4, P<.0001). The mortality rate was higher in HEN (RR 2.85; 95% CI 1.86-4.37, P<.0001) and in HON, (RR 1.6; 95% CI 1.04-2.52; P=.03), when compared to SIAD. The Charlson Comorbidity Index was lower in SIAD than in eunatraemic patients (P<.0001). 9/121(7.4%) patients died with plasma sodium <125 mmol/L and 4(3.3%) with plasma sodium <120 mmol/L. However, 69/121(57%) patients died with a plasma sodium above 133 mmol/L.

CONCLUSIONS

We confirmed higher all-cause mortality in hyponatraemia than in NN. Mortality was higher in SIAD than in normonatraemia and was not explained on the basis of co-morbidities. Mortality was higher in HON and HEN than in SIAD. Mortality rates reported for all-cause hyponatraemia in the medical literature are not applicable to SIAD.

Diabetes in Patients With Acromegaly

Acromegaly is a clinical syndrome which results from growth hormone excess. Uncontrolled acromegaly is associated with cardiovascular mortality, due to an excess of risk factors including diabetes mellitus, hypertension and cardiomegaly. Diabetes mellitus is a frequent complication of acromegaly with a prevalence of 12-37%. This review will provide an overview of a number of aspects of diabetes mellitus and glucose intolerance in acromegaly including the following: 1. Epidemiology and pathophysiology of abnormalities of glucose homeostasis 2. The impact of different management options for acromegaly on glucose homeostasis 3. The management options for diabetes mellitus in patients with acromegaly RECENT FINDINGS: Growth hormone and IGF-1 have complex effects on glucose metabolism. Insulin resistance, hyperinsulinaemia and increased gluconeogenesis combine to produce a metabolic milieu which leads to the development of diabetes in acromegaly. Treatment of acromegaly should ameliorate abnormalities of glucose metabolism, due to reversal of insulin resistance and a reduction in gluconeogenesis. Recent advances in medical therapy of acromegaly have varying impacts on glucose homeostasis. These adverse effects influence management choices in patients with acromegaly who also have diabetes mellitus or glucose intolerance. The underlying mechanisms of disorders of glucose metabolism in patients with acromegaly are complex. The aim of treatment of acromegaly is normalisation of GH/IGF-1 with reduction of co-morbidities. The choice of therapy for acromegaly should consider the impact of therapy on several factors including glucose metabolism.

The contribution of undiagnosed adrenal insufficiency to euvolaemic hyponatraemia: results of a large prospective single-centre study

OBJECTIVE

The syndrome of inappropriate antidiuresis (SIAD) is the commonest cause of hyponatraemia. Data on SIAD are mainly derived from retrospective studies, often with poor ascertainment of the minimum criteria for the correct diagnosis. Reliable data on the incidence of adrenal failure in SIAD are therefore unavailable. The aim of the study was to describe the aetiology of SIAD and in particular to define the prevalence of undiagnosed adrenal insufficiency.

DESIGN

Prospective, single centre, noninterventional, observational study of patients admitted to Beaumont Hospital with euvolaemic hyponatraemia (plasma sodium ≤ 130 mmol/l) between January 1st and October 1st 2015.

PATIENTS

A total of 1323 admissions with hyponatraemia were prospectively evaluated; 576 had euvolaemic hyponatraemia, with 573 (43·4%) initially classified as SIAD.

MAIN OUTCOME MEASURES

(i) Aetiology of SIAD, defined by diagnostic criteria; (ii) Incidence of adrenal insufficiency.

RESULTS

Central nervous system diseases were the commonest cause of SIAD (n = 148, 26%) followed by pulmonary diseases (n = 111, 19%), malignancy (n = 105, 18%) and drugs (n = 47, 8%). A total of 22 patients (3·8%), initially diagnosed as SIAD, were reclassified as secondary adrenal insufficiency on the basis of cortisol measurements and clinical presentation; 9/22 cases had undiagnosed hypopituitarism; 13/22 patients had secondary adrenal insufficiency due to exogenous steroid administration.

CONCLUSIONS

In a large, prospective and well-defined cohort of euvolaemic hyponatraemia, undiagnosed secondary adrenal insufficiency co-occurred in 3·8% of cases initially diagnosed as SIAD. Undiagnosed pituitary disease was responsible for 1·5% of cases presenting as euvolaemic hyponatraemia.