Sodium measurement: effects of differing sampling and analytical methods

Abstract

Background: It is recommended that children receiving intravenous fluids should have frequent biochemical monitoring, in some situations 4-6 hourly. Small changes in sodium must be detected, requiring very high precision from sodium analyses. Some children are monitored using venous blood analysed by indirect ion-selective electrode (ISE) interchangeably with capillary blood analysed by direct ISE. Our aim was to determine whether variability in sample collection together with variability in sodium measurement would lead to results which were unacceptable in the clinical setting. Methods: Fifty-seven adults had capillary and venous blood analysed for sodium using direct ISE and venous plasma analysed for sodium using indirect ISE. Results: Comparison of capillary blood analysed by direct ISE with venous plasma analysed by indirect ISE demonstrated wide scatter and poor correlation of results: r = 0.36, standard deviation (SD) of the differences 2.7 mmol/L and range of limits of agreement 10.6 mmol/L. Significant biases were observed comparing capillary blood sodium with venous plasma sodium ( P < 0.001), and comparing direct ISE with indirect ISE ( P < 0.001). Conclusions: Venous plasma using indirect ISE and capillary blood with direct ISE cannot be used interchangeably to detect small changes in plasma sodium concentrations. To avoid misinterpretation of results when monitoring sodium over short time periods, the use of single methods of sampling and analysis must be strongly encouraged.