The treatment of hyponatraemia is controversial because of the risk of causing central or extrapontine myelinolysis (EPM). Rapid correction with hypertonic saline to a low normal sodium level has its proponents; others feel that slow correction to below normal sodium values is preventative. Most investigators feel that overcorrection should be avoided. It is not known whether the magnitude of serum sodium change is more important than the actual rate of correction. We present three patients with hyponatraemia ranging from 103 to 105 mmol/l who were corrected slowly with normal saline, corrected quickly with hypertonic saline, or rapidly overcorrected with hypertonic saline. All became comatose and died; all had EPM with or without central pontine myelinolysis (CPM). The rate of correction, the solution used, or the magnitude of correction did not seem to protect against demyelination. In a review of 67 reported CPM cases since 1983, no patients documented as having CPM or EPM by radiological studies or necropsy were treated with water restriction only. A group of 27 hyponatraemic patients treated only with water restriction and 35 with diuretic cessation alone did not develop CPM or EPM. This may be a reasonable approach to patients with symptomatic hyponatraemia and normal renal function.