Pseudohypernatremia - evidence of a common problem

Around 25 % of blood samples recovered from patients admitted to intensive care and 8 % of those recovered from all other hospitalised patients would return a plasma sodium result 4-10 mmol higher than its true value if the method of analysis was indirect ion selective electrode (ISE). This is the headline finding of a recently published study that revisits a well-documented problem regarding measurement of plasma/serum sodium concentration. 

Although all plasma/serum sodium measurements are made using ISE nowadays, some methods require sample dilution prior to measurement and some do not, so a distinction is made between direct ISE (no dilution) methods, which are utilised in point-of-care analyzers - including blood gas analyzers, and indirect ISE (dilution) methods used widely in clinical laboratories.  

The two methods show generally good agreement as long as serum protein and lipid concentrations remain normal.  Direct ISE methods are unaffected by an abnormal change in protein and lipid concentrations but indirect ISE methods are affected. 

Increased plasma protein and/or lipid concentration can result in spuriously low plasma sodium (pseudohyponatremia) when measured by indirect ISE, whereas decreased plasma protein and/or lipid concentration causes spuriously high plasma sodium (pseudohypernatremia).  

Reduced plasma protein is a relatively common feature of acute/critical illness so the authors of this study sought to quantify the impact of abnormal plasma protein concentration on serum sodium measurement by indirect ISE in hospital patients.  They first studied 346 patients selected only on the basis of serum protein concentration, the object being to identify three approximately equal-sized groups each with different protein concentration.  

The first group comprised 117 patients with low protein concentration (<60g/L); the second comprised 105 patients with normal protein concentration (60-83g/L); and the third group comprised 124 patients with increased plasma protein concentration (> 83g/L). Blood was sampled from each of the 346 study patients for measurement of plasma sodium measurement by both indirect ISE (I) and direct ISE (D). 

For each participant the difference (I-D) between the two results was calculated.  By taking into account the analytical variability of the two methods and the biological variability of plasma sodium concentration, the authors determined that a difference  of > 4.0 mmol/L between indirect and direct method would be significant. They found as expected that the difference was progressively greater as the deviation from normal serum protein concentration increased. 

So the difference in sodium values (I-D) ranged from  plus 9 mmol/L for those with protein <40g/L to minus 10mmol/L for those with protein > 90g/L.  They determined the proportion of patients with serum protein concentration in each of 9 ranges (< 40g/L,  40-49g/L, 50-59g/L, 60-69 g/L etc up to >100g/L) who had I-D difference > 4.0 mmol/l. 

The authors applied these proportions to a large database of serum protein concentrations derived from 48,033 specimens from hospitalised patients, of which 2877 specimens were from patients admitted to intensive care. This analysis allowed their headline finding.  

Pseudohypernatraemia (due to low plasma protein concentration) was found to be much more common than pseudohyponatraemia (due to raised plasma protein concentration). 

In discussion of their findings the authors caution that their findings indicate that the use of indirect ISE methodology can commonly lead to clinically significant error in plasma sodium estimation. They call for the diagnostic industry to move to standardisation so that only direct ISE methods be used to measure plasma sodium.