Inaccuracy of Sodium Measurement in Patients with Severe Hypernatremia

We observed discordant sodium results from a patient with severe hypernatremia such that whole-blood analysis produced results up to 9.6 mmol/L higher than plasma sodium obtained on the same collection. We investigated this bias by comparing other patients' sodium results and performing comparisons of 3 blood gas and 2 chemistry analyzers.

First, the laboratory information system was queried for whole-blood sodium results >160 mmol/L, which were used for comparison against plasma results from the same collection. Second, whole blood was collected from a healthy donor, a portion of which was spiked with sodium chloride to generate 8 samples with target concentrations of 140 to 185 mmol/L. Whole-blood sodium was measured in duplicate on the ABL90, RAPIDPoint 500, and GEM 4000. Plasma sodium was then measured in duplicate on the Architect c8000 and Cobas c702. Finally, plasma was injected on the blood gas analyzers to measure sodium in singleton.

Overall, 53 paired results from patients showed a significant positive bias on the ABL90 relative to Vitros when sodium was >160 mmol/L. The magnitude of difference was insignificant within the reference range but increased proportionately with concentration. The magnitude and pattern of positive bias in ABL90 sodium results were consistent with the observation in patient results.

In severe hypernatremia, sodium results produced by blood gas and plasma analyzers can differ significantly.