The effect of hyperglycemia is well known for its lowering of serum sodium levels. The most commonly used correction factor is a 1.6 mEq per L (1.6 mmol per L) decrease in serum sodium for every 100 mg per dL (5.6 mmol per L) increase in glucose concentration. Hillier and associates challenged this correction factor in a study that monitored serum sodium levels in a group of otherwise healthy persons who were rendered acutely insulin deficient.

Healthy young volunteers were eligible for the study if they had no personal or family history of diabetes mellitus and were not taking any medications. After an overnight fast, baseline serum glucose and sodium levels were measured. Subsequently, somatostatin was infused to suppress insulin secretion, along with a solution of 20 percent dextrose to increase the glucose concentration to more than 600 mg per dL (33.3 mmol per L) within one hour. Once this glucose level was attained, the dextrose infusion was stopped and regular insulin was given (a six-unit bolus and then six units an hour) until the glucose concentration was less than 140 mg per dL (7.8 mmol per L). Serum sodium and plasma glucose levels were measured three times in three subjects during induction of hyperglycemia. These measurements were obtained in all subjects every 10 minutes once the glucose infusion was stopped and their glucose levels had returned to normal.

A group of six volunteers, consisting of five men and one woman, completed the study. Serum sodium levels decreased in all volunteers as their glucose levels increased. When placed on a simple straight line regression, the average slope was −2.4 (± 0.3) mEq per L sodium per 100 mg per dL glucose. This factor is significantly greater than the conventional correction factor. However, the decrease in serum sodium did not actually appear to be a linear value, particularly in volunteers whose glucose concentration was greater than 440 mg per dL (24.4 mmol per L). At this level, the slope was about −4.0, but it dropped to −1.6 when glucose concentration dropped below 440 mg per dL. The 2.4 correction factor was slightly inaccurate in the 300 to 500 mg per dL (16.7 to 27.8 mmol per L) range, but was accurate for values in the more clinically important range of 500 mg per dL or greater. The conventional correction factor underestimated sodium concentration in most cases when the glucose concentration was more than 300 mg per dL, and underestimated almost all measurements when the glucose concentration was more than 500 mg per dL.

The authors conclude that the conventional correction factor used to correct serum sodium in patients with hyperglycemia is inaccurate and leads to serious underestimation of serum sodium values in association with glucose levels higher than 500 mg per dL. Therefore, the authors recommend using 2.4 as the correction factor to prevent such underestimations, particularly in patients with severe hyperglycemia who require more of a correction.