Katz vs Hillier: which sodium correction factor for hyperglycemia?

Where the Katz 1.6 correction factor comes from

The Katz correction factor of 1.6 mmol/L of sodium per 100 mg/dL of glucose comes from a short letter by M. A. Katz in the New England Journal of Medicine in October 1973, titled "Hyperglycemia-induced hyponatremia: calculation of expected serum sodium depression". As that title says, the figure was calculated rather than measured in patients. The physiology behind it is described in the 2009 American Diabetes Association consensus statement on hyperglycemic crises: when insulin is lacking, the admission serum sodium is usually low because of the osmotic flux of water from the intracellular to the extracellular space in the presence of hyperglycemia. Glucose that stays outside cells pulls water out of them, and that water dilutes the sodium. Adding 1.6 mmol/L for every 100 mg/dL of glucose above 100 mg/dL estimates the sodium the patient would have once that water moved back. For 26 years the factor was used widely, although, as Hillier's team later pointed out, no controlled experiment had tested it.

In the corrected sodium calculator, the Katz result is the first of the two values shown.

What Hillier measured in 1999

The Hillier correction factor of 2.4 comes from an experiment published by Hillier, Abbott and Barrett of the University of Virginia in the American Journal of Medicine in April 1999. They infused somatostatin into 6 healthy subjects to block their own insulin secretion, then raised plasma glucose above 600 mg/dL within one hour with a dextrose infusion. The infusion was stopped and insulin given until glucose fell back to 140 mg/dL, with glucose and sodium measured every 10 minutes throughout. Overall, sodium fell by a mean of 2.4 mmol/L for every 100 mg/dL rise in glucose, significantly more than the 1.6 factor (P = 0.02). The relationship was also nonlinear, most clearly above 400 mg/dL. Up to 400 mg/dL the standard 1.6 correction worked well, but above 400 mg/dL a factor of 4.0 was better. The authors concluded that 2.4 is a better overall estimate of the association than the usual 1.6.

Katz vs Hillier: which correction factor to use

Neither the Katz nor the Hillier factor is exact for every patient, and the gap between them only matters at high glucose. Below 400 mg/dL (22.2 mmol/L), Hillier's own data showed that 1.6 worked well, and the two corrections differ by 2.4 mmol/L at most. Above 400 mg/dL, 2.4 is closer to what Hillier measured, and even 2.4 may under-correct because the fall steepened toward a factor of 4.0. Guidelines have kept the older figure: the 2009 ADA consensus statement corrects sodium by adding 1.6 per 100 mg/dL of glucose above 100 mg/dL, and the 2024 consensus report by the ADA, EASD, JBDS, AACE and the Diabetes Technology Society states that a fall in glucose of 5.6 mmol/L (100 mg/dL) results in a 1.6 mmol/L rise in sodium. A practical answer is to look at both, which is why the calculator shows them side by side, and to say which factor you used.

Values are mmol/L of sodium added to the measured sodium. A gap of 4 to 7 mmol/L is enough to move a result across the 145 mmol/L line, so at these glucose levels the factor you pick can change how the result reads.

Reading corrected sodium in DKA and HHS

In diabetic ketoacidosis (DKA) and the hyperglycemic hyperosmolar state (HHS), corrected sodium helps judge the water deficit hidden behind a low or normal measured sodium. The 2009 ADA consensus statement says that an increased or even normal serum sodium in the presence of hyperglycemia indicates a profound loss of water, and corrects sodium to assess the severity of the sodium and water deficit. A corrected sodium in the normal range means the low measured value is explained by glucose; a corrected sodium above 145 mmol/L means water loss outweighs sodium loss; one still below 135 mmol/L means a sodium deficit or water excess remains. The same arithmetic predicts what happens next: the 2024 consensus report explains that as glucose falls, water shifts back into cells and measured sodium may rise, by 1.6 mmol/L for each 100 mg/dL fall in glucose. A rising measured sodium during treatment is therefore expected when the corrected sodium was already higher.

Two values are easy to mix up. Effective serum osmolality, which the 2009 ADA statement uses to define HHS (above 320 mOsm/kg, with glucose above 600 mg/dL, arterial pH above 7.30 and bicarbonate above 18 mmol/L), is calculated from the measured sodium: 2 × measured Na + glucose (mg/dL) ÷ 18. Putting the corrected sodium into that formula would count the glucose twice. The 2024 report also updated the diagnostic criteria; for DKA it now accepts a glucose of 200 mg/dL (11.1 mmol/L) or more, or a known history of diabetes. For HHS, the 2024 report adds that the measured serum sodium should not fall by more than 10 mmol/L in 24 hours, so serial measured values matter as much as the corrected one. Run each new pair of values through the corrected sodium calculator to follow the corrected sodium over time.

Worked examples with both correction factors

Four sets of results show when the choice between the Katz and Hillier factors changes the reading and when it does not. In the first, a measured sodium of 128 mmol/L with a glucose of 600 mg/dL (33.3 mmol/L) has 500 mg/dL of glucose above 100, so Katz adds 1.6 × 5 = 8.0 and Hillier adds 2.4 × 5 = 12.0, giving 136.0 and 140.0 mmol/L. Both are normal, so the low measured sodium is explained by the glucose. In the second, the same sodium of 128 with a glucose of 900 mg/dL gives 128 + 12.8 = 140.8 by Katz but 128 + 19.2 = 147.2 by Hillier, and the two factors disagree on whether there is a water deficit. In the third, a sodium of 140 with a glucose of 1000 mg/dL corrects to 154.4 (Katz) or 161.6 (Hillier): both high, a large water deficit by either factor.

In the third case, the effective osmolality uses the measured sodium: 2 × 140 + 1000 ÷ 18 = 335.6 mOsm/kg, above the 320 mOsm/kg in the 2009 HHS criteria. In the fourth, a sodium of 124 with a glucose of 300 mg/dL corrects only to 127.2 or 128.8, so most of the hyponatremia is not due to glucose. Enter any of these in the corrected sodium calculator to check the numbers.