Background: Electrolyte values are measured in most critically ill intensive care unit (ICU) patients using both an arterial blood gas analyzer (ABG) and a central laboratory auto-analyzer (AA). The aim of the present study was to investigate whether electrolyte levels assessed using an ABG and an AA were equivalent; data on sodium and potassium ion concentrations were examined.
Methods: We retrospectively studied patients hospitalized in the ICU between July and August 2011. Of 1,105 test samples, we identified 84 instances of simultaneous sampling of arterial and venous blood, where both Na+ and K+ levels were measured using a pHOx Stat Profile Plus L blood gas analyzer (Nova Biomedical, Waltham MA, USA) and a Roche Modular P autoanalyzer (Roche Diagnostics, Mannheim, Germany). Statistical measures employed to compare the data included Spearman's correlation coefficients, paired Student's t-tests, Deming regression analysis, and Bland-Altman plots.
Results: The mean sodium concentration was 138.1 mmol/L (SD 10.2 mmol/L) using the ABG and 143.0 mmol/L (SD 10.5) using the AA (p < 0.001). The mean potassium level was 3.5 mmol/L (SD 0.9 mmol/L) using the ABG and 3.7 mmol/L (SD 1.0 mmol/L) using the AA (p < 0.001). The extent of inter-analyzer agreement was unacceptable for both K+ and Na+, with biases of 0.150-0.352 and -0.97-10.05 respectively; the associated correlation coefficients were 0.88 and 0.90.
Conclusions: We conclude that the ABG and AA do not yield equivalent Na+ and K+ data. Concordance between ABG and AA should be established prior to introduction of new ABG systems.