An alcoholic patient presented with profound hyponatremia (serum sodium concentration, 96 mEq/L) caused by the combined effects of a thiazide diuretic, serotonin reuptake inhibitor, beer potomania, and hypovolemia. A computed tomographic scan of the brain was indistinguishable from one obtained 3 weeks earlier when he was normonatremic. Concurrent administration of 3% saline solution and desmopressin controlled the rate of correction to an average of 6 mEq/L daily and resulted in full neurologic recovery without evidence of osmotic demyelination. This case illustrates the value of controlled correction of profound hyponatremia.
This is a very tough situation. What do you do? No matter what you give the patient you have no control over how fast the serum sodium rises. The patient, depending on his volume status, may have volume signals for vasopressin which could be shut off by the administration of volume in any form, resulting in a free water diuresis and uncontrollable rise in the serum sodium. The solution to this case was to make sure, while 3% saline was given, to maintain sufficient antidiuretic effect via administration of desmopressin to prevent that outcome.
I first read about the use of desmopressin to avoid over correction of hyponatremia in UptoDate several years ago. At that time it was advocated as a reactionary move---something to do when you saw the sodium rising too fast. Later on I blogged about a paper describing the use of desmopressin as a preemptive strike in certain clinical circumstances. This AJKD paper is the first I've seen to advocate desmopressin as part of the initial therapy, but it makes sense. Although the authors didn't give the precise recipe for desmopressin, the full text is worth the read. This paragraph from the body of the paper explains why the rise of serum sodium with any therapy would be impossible to predict or control:
Impaired diluting ability due to a thiazide diuretic and syndrome of inappropriate secretion of antidiuretic hormone caused by an SSRI will resolve soon after these medications are cleared,  ,  ,  and our patient most likely stopped taking these medications many hours before admission. Similarly, vasopressin secretion in response to hypovolemia ceases when the volume deficit is corrected using saline. Our patient had a very low urine sodium concentration (Table 1), consistent with hypovolemia. Transient vasopressin secretion related to alcohol withdrawal and the stress of a seizure will resolve spontaneously. Our patient had a history of alcohol withdrawal seizures and was delirious despite an undetectable blood alcohol level (Table 1). When nonosmotic stimuli for vasopressin secretion are eliminated, hypothalamic vasopressin-secreting neurons respond to the low plasma sodium concentration, and plasma vasopressin levels rapidly become undetectable. This results in excretion of up to 1 L/h of maximally dilute urine that will increase serum sodium concentration by greater than 2 mEq/L/h.
H/T to DB's MedRants.