A problem in sodium measurement addressed

The concentration of sodium (Na) in blood serum/plasma is determined by ion-specific electrode (ISE) technology that measures the physiologically important activity of sodium ions present in the aqueous (water) phase of plasma. 

Two methods are available: direct ISE and indirect ISE. Direct ISE, which is employed in blood gas and other point-of-care analyzers, involves presentation of an undiluted whole-blood sample to the measuring electrode. By contrast indirect ISE, the method of choice for central laboratory analyzers, involves presentation of a prediluted serum or plasma sample to the measuring electrode. 

The accuracy of indirect ISE Na measurement depends crucially on the assumption that the volume of serum/plasma occupied by water is ~93 % and the remaining ~7 % is occupied by lipids and proteins (the non-water fraction). 

Indirect ISE sodium measurement can give rise to spuriously decreased serum/plasma sodium concentration (pseudohyponatremia) if lipids and/or protein are significantly raised because under these circumstances the non-water fraction is >7 %, the water fraction is <93 % and the diluted sample presented to the electrode contains fewer sodium ions than would be the case if the sample had normal proportion of water (93 %). 

Conversely, if serum lipid and/or protein concentration is markedly reduced, then non-water fraction is <7 % and water fraction is >93 %, resulting in spuriously raised sodium (pseudohypernatremia). Direct ISE is unaffected by abnormality in lipid and protein concentration simply because it involves no sample predilution.

Although the potential for spurious sodium results (pseudohyponatremia and pseudohypernatremia) associated with use of indirect ISE has been known since the 1980s, there has been realization in recent years that pseudohypernatremia is a particular problem for one patient group, the critically ill. 

Reduced plasma albumin and therefore reduced plasma total protein (hypoproteinemia) is a common feature of critical illness. The relatively high prevalence of hypoproteinemia and consequent pseudohypernatremia among the critically ill has persuaded some to argue that for this patient group at least indirect-ISE methods should not be used to measure plasma/serum sodium. 

A potential alternative solution to the problem, which some, including the authors of a recently published US study, have investigated is development of a correction factor based on patient serum protein or serum albumin concentration that could be applied to indirect ISE sodium results in order to predict a true (direct ISE) value.

For this study researchers interrogated their hospital laboratory database for results relating to patients admitted to critical care units over a 2-year period. They were looking for occasions of patient’s blood being submitted for analysis at both the central laboratory analyzer (which measures Na by indirect ISE) and blood gas analyzer (which measures Na by direct ISE). 

So long as the interval between the two analyses was less than 20 minutes, the patient’s paired results were considered concurrent and abstracted for study. In the event they abstracted 774 paired indirect ISE and direct ISE sodium results. 

For each pair of results, total protein, albumin, glucose and tCO₂ values generated by the central laboratory analyzer, as well as pH and glucose generated by the blood gas analyzer, were also abstracted. 

For each pair the difference between indirect and direct sodium (∆Na) was calculated; the mean ∆Na was 2.3 ± 2.2 mmol/L. Investigators confirmed that both serum albumin and serum protein concentration vary inversely with ∆Na, although the relationship was found to be linear only for total protein, so further statistical analysis was confined to the relationship between total protein concentration and ∆Na. 

Multiple linear regression analysis revealed that for every 1 g/dL increase in total protein concentration, ∆Na decreased by 0.69 ± 0.06 mmol/L so that the authors of the study were able to advise that in order to correct a patient’s indirect ISE sodium result for change in total protein, one must add 0.7 mmol/L for each 1 g/dL increase in total protein concentration and subtract 0.7 mmol/L for each 1 g/dL decrease in total protein.

The study report includes discussion of the limitations of previous studies aimed at determining a reliable correction factor, and how knowledge of these limitations informed the study design and choice of more sophisticated statistical analysis used to generate the advised correction factor.