Objectives: To test whether a drop in effective plasma osmolality (P(Eff osm); 2 x plasma sodium [P(Na)] + plasma glucose concentrations) during therapy for diabetic ketoacidosis (DKA) is associated with an increased risk of cerebral edema (CE), and whether the development of hypernatremia to prevent a drop in the P(Eff osm) is dangerous.
Study design: This study is a retrospective comparison of a CE group (n = 12) and non-CE groups with hypernatremia (n = 44) and without hypernatremia (n = 13).
Results: The development of CE (at 6.8 +/- 1.5 hours) was associated with a drop in P(Eff osm) from 304 +/- 5 to 290 +/- 5 mOsm/kg (P < .001). Control patients did not show this drop in P(Eff osm) at 4 hours (1 +/- 2 and 2 +/- 2 vs -9 +/- 2 mOsm/kg; P < .01), because of a larger rise in P(Na) and/or a smaller drop in plasma glucose. During this period, the CE group received more near-isotonic fluids (69 +/- 9 vs 35 +/- 2 and 27 +/- 3 mL/kg; P < .001). The CE group had a higher mortality (3/12 vs 0/57; P = .003), and more neurologic sequelae (5/12 vs 1/57; P < .001).
Conclusions: CE during therapy for DKA was associated with a drop in P(Eff osm). An adequate rise in P(Na) may be needed to prevent this drop in P(Eff osm).