Diabetes insipidus and the failed osmostat

Killing ADH-loop gain breaks the osmostat: an imposed water deficit drives osmolality up and it never returns. At a 3 L deficit the engine holds osmolality at 309 mOsm/kg (a hypernatremia surrogate) versus 287 for the intact loop, which corrects every load. Grade [V]; DI hyperosmolality anchor [L]; absolute Na [O].

Diabetes insipidus sets the ADH/thirst gain to zero on the osmostat. Across graded water deficits the diseased loop fails to recover — osmolality climbs 290→317 mOsm/kg and tracks the deficit — while the intact loop returns to the 287 setpoint within 1 mOsm. The contrast, not a tuned number, is the result.

No antidiuresis, no recovery

In diabetes insipidus the antidiuretic response is absent (central) or unheeded (nephrogenic). On the osmostat this is the loop gain going to zero: the controller can no longer retain free water, so an osmotic load is not corrected.

Hyperosmolality that tracks the deficit

Imposing water deficits of 0.5–4 L, the gain-zero loop lets osmolality rise monotonically 290→317 mOsm/kg and stay there, crossing the hyperosmolality band (≥295, a hypernatremia surrogate, Na>145), while the intact loop corrects every load back to 287 mOsm/kg (within 0.2). Each diseased value exceeds its intact counterpart by far more than the loop's own residual — a mechanistic separation rather than a fitted offset. The direction and contrast are reproduced [V]; mapping osmolality to an absolute plasma sodium needs a renal free-water-clearance calibration and stays [O].