1. Osmotically stimulated thirst and vasopressin release were studied during infusions of hypertonic sodium chloride and hypertonic d-glucose in euglycaemic clamped diabetic patients and healthy controls.
2. Infusion of hypertonic sodium chloride caused similar elevations of plasma osmolality in diabetic patients (288.0 ± 1.0 to 304.1 ± 1.6 mosmol/kg, mean ± sem, P < 0.001) and controls (288.6 ± 0.9 to 305.7 ± 0.6 mosmol/kg, P < 0.001), accompanied by progressive increases in plasma vasopressin (diabetic patients, 0.9 ± 0.3 to 7.7 ± 1.5 pmol/l, P < 0.001; controls 0.5 ± 0.1 to 6.5 ± 1.0 pmol/l, P < 0.001) and thirst ratings (diabetic patients 1.0 ± 0.2 to 7.1 ± 0.5 cm, P < 0.001; controls 1.8 ± 0.4 to 8.0 ± 0.5 cm, P < 0.001) in both groups.
3. Drinking rapidly abolished thirst and vasopressin secretion before major changes in plasma osmolality occurred in both diabetic patients and healthy controls.
4. There were close and significant correlations between plasma vasopressin and plasma osmolality (diabetic patients, r = + 0.89, controls r = + 0.93) and between thirst and plasma osmolality (diabetic patients r = +0.95, controls r = +0.97) in both diabetic patients and healthy controls during hypertonic saline infusion.
5. Hypertonic d-glucose infusion caused similar elevations in blood glucose in diabetic patients (4.0 ± 0.2 to 20.1 ± 1.2 mmol/l, P < 0.001) and healthy controls (4.3 ± 0.1 to 19.3 ± 1.2 mmol/l, P < 0.001) but did not change plasma vasopressin or thirst ratings. There was no correlation between plasma osmolality and either thirst or plasma vasopressin during hypertonic d-glucose infusion.
6. The characteristics of osmoregulated thirst and vasopressin release are similar in health and diabetes mellitus. As hyperglycaemia was not dipsogenic, however, the thirst of poorly controlled diabetes mellitus may be due to hypovolaemia secondary to polyuria rather than hyperosmolality due to elevated blood glucose concentrations.
