Diuretics are recommended as first-line antihypertensive treatment in elderly patients. Although attention is usually paid to prevent hypokalaemia with diuretic therapy, risk of hyponatraemia is often ignored. We performed this study to characterise hypertensive patients at increased risk to develop hyponatraemia. We reviewed charts of hypertensive patients hospitalised in Chaim Sheba Medical Center for hyponatraemia from 1990 to 1997. Patients with other causes of hyponatraemia were excluded. The General Practice Maccabi database was used to estimate age and sex distribution of patients prescribed diuretics for hypertension. We identified 180 hypertensive patients (149 F, 31 M; mean age 76.4 +/- 9.2 years) hospitalised because of hyponatraemia. Across all age groups, odds ratio (OR) to develop hyponatraemia was three times higher for women vs men (OR 3.10, 95% confidence interval (CI): 2.07-4.67). One hundred and sixty-two patients (90%) were older than 65 years. Patients of both sexes older than 65 years were 10 times (and if they were older than 75 years 16 times) more likely to develop hyponatraemia than those younger than 65 years (OR 9.87, 95%, CI: 5.93-16.64). Most patients (74.5%) used a thiazide-based diuretic; only 10% used a low dose (<25 mg/day). In 37% of patients diuretics were used for more than 1 year before hyponatraemia developed. Diuretic-induced hyponatraemia may be insidious and appear even after prolonged diuretics use. Elderly women seem to be at particularly high risk. In this population diuretic use should be associated with close monitoring of sodium and potassium levels.