We evaluated in 110 consecutive elderly hyponatremic patients the value of traditional clinical and biochemical data and the place of a test infusion of 2 liters isotonic saline over 24 hours, in establishing the etiology of the hyponatremia. The causes of hyponatremia were as follows: 31% SIADH patients, 23% patients with hyponatremia due to diuretics, 18% potomania patients, 15% salt depleted patients, 5% salt depleted SIADH patients, 5% patients with a salt loosing syndrome and 3% patients with hyponatremia of unknown origin. Several salt depleted (SD) and SIADH patients could be confounded. Usually, adults with SIADH show plasma uric acid values <4 mg/dL. In our elderly population, 41% of SD patients presented plasma uric acid <4 mg/dL, while 27% of SIADH patients showed plasma uric acid >4 mg/dL. Eighty-two percent of SD patients appeared to have plasma urea levels >30 mg/dL, but this was also the case in 21% of SIADH patients. Twenty-nine of the SD patients presented a urinary sodium >30 mEq/L, but all had fractional sodium excretion (FENa) lower than 0.5%. However, in SIADH, 42% of the patients presented also FENa <0.5%. Fractional excretion of urea (FE urea) below 50% was encountered in 82% of SD patients and FE urea above 50% in only 52% of the SIADH patients. Plasma renin and aldosterone values were poorly discriminative. A test infusion with 2 liters isotonic saline over 24 hours allowed a correct classification of all the patients. In about 2/3 of the population, administration of isotonic saline could be considered as useful (SD, most diuretic patients, potomania patients, salt loosing syndrome patients and some SD SIADH patients). A plasma sodium (PNa) increase of at least 5 mEq/L 24 hours after saline infusion has been suggested as highly suggestive of SD. Nevertheless, 29% of our SD patients did not increase their PNa level by 5 mEq/L or more, while 30% of our SIADH patients did. PNa improved after 2 liters isotonic saline over 24 hours in 90 patients (85%) as opposed to 12 others (9 SIADH and 3 diuretic patients), decreasing their plasma sodium. The isotonic saline infusion test, only allows a reliable classification of hyponatremia, as far as both PNa and sodium excretion were taken into account. In the SIADH group, 6 patients (5%) presented initially manifest solute depletion and retained the 2 liters isotonic saline before developping inappropriate natriuresis. Six patients showed a transient salt loosing syndrome with high fractional potassium excretion (FEK) and high calciuria, which differentiates them from thiazide patients presenting also high FEK, but low calciuria. These patients were also polyuric at admission. The saline infusion was well tolerated in all but 2 patients, developing mild pulmonary congestion at the end of the test infusion.