Objective To determine the characteristics, causes and outcome of severe hyponatraemia (< 125 mmol/l) in hospitalized patients, and to identify mortality predictors.
Design Prospective case controlled study of sequentially presenting patients with a serum sodium (Na) < 125 mmol/l.
Patients and methods One hundred and four hyponatraemic and 104 randomly chosen normonatraemic (Na > 135 mmol/l) adult patients were studied. We measured hospital mortality and days in hospital, diagnoses, drug history and cause of hyponatraemia. Na was recorded at admission, as well as the closest level measured before death or discharge. In addition, the lowest Na was recorded (if this was not at admission).
Results Hyponatraemic patients were older (mean age ± 1 SD 69 ± 14 years) than controls (61 ± 16 years, P < 0·001), but of similar sex ratio. On admission, Na in the hyponatraemic group was 125 ± 7 mmol/l compared with 139 ± 3 mmol/l in controls (P < 0·0001), but fell to 120 ± 4 mmol/l before rising at discharge to 131 ± 7 mmol/l (all changes P < 0·001). Overall mortality was 27% in hyponatraemic patients compared with 9% in controls (P = 0·009), and length of admission was also greater (16 ± 12 vs. 13 ± 11 days, P < 0·005). Mortality was greater in patients whose Na levels fell during admission (34%vs. 16%, P < 0·05), and these patients appeared to have an excess of diuretic-induced and possibly iatrogenic hyponatraemia.
Conclusions Severe hyponatraemia in hospital patients is associated with prolonged admissions and significantly increased mortality compared with normonatraemic patients. A particular group at high risk of death are those whose Na levels fall after admission. They may represent a ‘sicker’ group, and deserve increased monitoring and surveillance.