Endocrine Abstracts

Introduction: Hyponatraemia (serum sodium <136 mmol/l) is associated with significant morbidity and mortality. International guidelines suggest a clear algorithm for investigation, inclusive of measurements of paired urine sodium and osmolality, thyroid function tests and a morning cortisol. The aim of this study was to prospectively investigate the assessment, management and clinical outcomes associated with hyponatraemia on hospital admission.

Methods: This prospective study was conducted from 03/09/2018-03/10/2018, and follow-up data was collected six months thereafter. Information was garnered on hyponatraemic admissions through a combined review of patient charts and the hospital’s laboratory database.

Results: Of the 418 patients admitted, 75 (18%, 35 male, 40 female) had hyponatraemia, with a mean age of 74 years (s.d. = 14). Eleven were excluded on the basis of a recent surgical or oncological admission. 63 (84%) had mild (130–135 mmol/l), 9 (12%) had moderate (125–129 mmol/l) and 3 (4%) had severe (<125 mmol/l) hyponatraemia. Only 4 (5%) patients had measurements of paired serum and urine osmolality and sodium, 19 (25%) of thyroid function, and 1 (1%) of morning cortisol. Only 9 (12%) were assessed by a consultant endocrinologist. 47 (63%) were taking a culprit medication (ACE inhibitor, ARB, diuretic, SSRI, SNRI, tricyclic antidepressant) on admission, 15 (32%) of which were ceased. The mean length of hospital stay was 7 days for mild, 9 days for moderate and 16 days for severe cases, and there were 2 in-hospital mortalities. Of the 73 surviving patients, 23 (31%) did not have a sodium measurement performed at discharge, and 27 (37%) were discharged with persistent hyponatraemia. Over a 6 month follow-up period, 26/73 (36%) of hyponatraemic cases were readmitted, compared with 100/332 (30%) of normonatraemic patients. Over the same period, hyponatraemic admissions had a mortality rate of 16% (12/73), while normonatraemic admissions had a rate of 4% (13/332). The difference in proportions is significant, χ²(1, N = 405) = 16.2, P < 0.001.

Conclusions: Hyponatraemia was a highly prevalent condition on admission which was largely under investigated; laboratory tests recommended by current expert guidelines were underutilised, and specialist advice was rarely sought. Management was also suboptimal; significant proportions were discharged without measurement or correction of serum sodium concentrations. Hyponatraemia was ultimately associated with a 4.5-fold excess in mortality at six months post-discharge. These findings emphasise the need for the development of Irish guidelines and the introduction of electronic alert systems to improve hospital practice.