RESUMO
INTRODUCTION: The determination of the electrolytes sodium and potassium is essential in critical care. In daily clinical practice, both the blood gas analyzer (ABG) and the laboratory autoanalyzer (AA) are generally applied. However, there is still uncertainty regarding the convergence of the prementioned assays, and data about the comparability dependent on the pH value are still lacking. MATERIALS AND METHODS: One hundred samples from intensive care unit patients with a range in pH values between 7.20 and 7.49 were evaluated in this retrospective cohort study. All patients suffered an infarct-related cardiogenic shock and were intubated and not under therapeutical hypothermia at the time of blood collection. We used scatter plots to compare different distributions of sodium and potassium values between the methods. Comparability of the analyses was assessed using the Bland-Altmann approach, and intraclass correlations (ICC) as estimates of interrater reliability were calculated. RESULTS: The mean potassium level measured on ABG was 4.33 mmol/L (SD 0.48 mmol/L), and the value obtained using the AA was 4.40 mmol/L (SD 0.55 mmol/L). A Bland-Altman comparison for total potassium measurements revealed that the limits of agreement were small (-0.241 to 0.391 mmol/L). Total ICC displayed a very good correlation of 0.949. For sodium, we found average values of 140 mmol/L (SD 5.20 mmol/L) in the AA and 140 mmol/L (SD 5.80 mmol/L) in the ABG assessment. Contrarily, the Bland-Altman comparison for sodium displayed that the 95% limits of agreement were very wide (-5.99 to 6.59 mmol/L) for total measurements as well as in every pH subgroup. Total ICC only reached a value of 0.830. CONCLUSION: Data from our single-center study indicate that urgent and vital decisions based on potassium measurements can be made by trusting the value obtained on the ABG machine irrespective of pH values.