Background We compared the effect of haemolysis in sodium measurement using indirect and direct ion-selective electrodes to test the hypothesis that haemolytic effect on sodium would be greater with indirect ion-selective electrode due to electrolyte exclusion effect from released intracellular proteins. Methods Plasma lithium heparin samples ( n = 36) from four volunteers were prepared to give a range of haemolytic indices (H-indices). Samples were analysed for sodium by indirect ion-selective electrode, H-index and total protein on an Abbott Architect c16000 and sodium by direct ion-selective electrode on a Siemens RAPIDPoint 500. Percentage changes in sodium in paired direct and indirect ion-selective electrode values were compared. Results Abbott H-index, which represents haemoglobin concentration in g/L, correlated with percentage negative change in sodium by direct ion-selective electrode (ρ 0.995, P < 0.001) and indirect ion-selective electrode (ρ 0.991, P < 0.001). Percentage negative change was less when sodium was measured by direct ion-selective electrode compared to indirect ion-selective electrode (Wilcoxon signed-rank Z = 3.46, P = 0.01). The difference in percentage change in sodium between direct ion-selective electrode and indirect ion-selective electrode correlated with total protein (ρ 0.751, P < 0.001). The negative bias in sodium results exceeded the reference change value of 2.2% at an H-index of 8.31 for indirect ion-selective electrode and 9.26 for direct ion-selective electrode. Conclusion Haemolysis causes negative influence with sodium measured by both indirect and direct ion-selective electrode due to a dilutional hyponatremia. The additional interference in indirect ion-selective electrode is due to the electrolyte exclusion effect but this is unlikely to be clinically significant as it is small in magnitude.

Background Haemolysis has been reported as having a positive, negative or no effect on plasma sodium (PNa) and chloride (PCl). We investigated the haemoltytic effect of different haemolysis protocols on PNa and PCl using modelling and laboratory experiments. Methods In a modelling experiment, percentage change and recovery due to dilution in routinely ( in vitro) haemolysed samples were compared against shear stress haemolysis and samples spiked with haemolysate from whole blood freeze–thaw, packed cells freeze–thaw and osmotic shock protocols. The results were compared against a control base pool. Additionally, for the osmotic shock method, results were compared against saline- and deionised water (DIW)-spiked controls. In a laboratory experiment, percentage change and recovery were similarly compared using haemolysate from whole blood freeze–thaw and osmotic shock protocols. PNa, PCl and H-index were measured on the Abbott Architect and haemoglobin on the Sysmex XN-9000. Results In the modelling experiment, the percentage decrease in PNa and PCl was similar in in vitro haemolysis, shear stress haemolysis, whole blood freeze–thaw haemolysis and packed cells freeze–thaw haemolysis and this was lower compared to the osmotic shock method. In the laboratory experiment, the change in PNa compared to the base pool was less ( p < 0.001) per unit increase in H-index in the freeze–thaw method (−0.33 mmol, 95% CI −0.35 to −0.31) compared to the osmotic shock method (−0.65 mmol, 95% CI −0.66 to −0.64). PCl did not change with haemolysis in the freeze–thaw method and changed by −0.21 ± 0.01 mmol per unit increase in the H-index in the osmotic shock method. Recovery of PNa and PCl increased with increasing H-index in both methods. Conclusion The osmotic shock protocol is inappropriate for haemolysis studies because of dilution with DIW used for cell lysis. Recovery calculations may incorrectly compensate for genuine dilution caused by haemolysis.