Objectives: This study was conducted to evaluate the measurement bias, correlation, and limits of agreement between arterial and venous blood in electrolyte testing.
Methods: A prospective cross-sectional study was conducted at Gia Dinh People’s Hospital from June to December 2025. Hospitalized patients with simultaneous orders for ABG and venous biochemistry tests (using lithium heparin tubes) were enrolled. Paired arterial and venous blood samples were collected concurrently and analyzed using the RapidPoint 500 blood gas analyzer (direct ISE) and the AU5800 biochemistry analyzer (indirect ISE), respectively. Statistical analysis included paired t-tests, Pearson correlation coefficients, and Bland-Altman plots.
Results: 347 paired arterial and venous blood samples. The mean electrolyte concentrations in arterial blood were consistently lower than those in venous blood. The mean differences (arterial - venous) were -3.11 mmol/L for Na⁺, -0.45 mmol/L for K⁺, and -2.58 mmol/L for Cl⁻, all of which were statistically significant (p < 0.001). Pearson correlation analysis demonstrated strong linear correlations between the two sample types (Na⁺: r = 0.857; K⁺: r = 0.887; Cl⁻: r = 0.853; all p < 0.001). However, Bland-Altman analysis revealed wide limits of agreement, particularly for Na⁺ and Cl⁻. When compared with Total Allowable Error (TEa) criteria, the proportion of samples exceeding acceptable clinical limits was substantial (Na⁺: 37.8%; K⁺: 33.4%; Cl⁻: 82.7%), indicating limited agreement between the two measurement methods.
Conclusions: This study demonstrated a significant difference between electrolyte results obtained from arterial blood (direct ISE) and venous blood (indirect ISE). The observed bias exceeded the Total Allowable Error, particularly for sodium (Na⁺) and chloride (Cl⁻), indicating that the two methods are not interchangeable in clinical practice. However, potassium (K⁺) results obtained from arterial blood gas analysis may be used for reference in emergency situations when venous sampling is not feasible, but they should be confirmed by standard biochemical testing under routine conditions.