Among 1,870 critically ill patients with acute kidney injury, sodium and potassium from arterial blood-gas analysers showed relatively good agreement with central laboratory values, with potassium having the narrowest limits. Clinically meaningful discrepancies occurred in 7.0% of sodium, 14.1% of potassium and 31.4% of chloride measurements, so rapid bedside results should support rather than replace reference testing.
Key findings
- Mean bias (95% limits of agreement) was -0.68 mmol/L (-11.41 to 10.06) for sodium, -0.17 (-1.46 to 1.12) for potassium and 7.18 (-4.15 to 18.51) for chloride. Correlations were 0.736, 0.642 and 0.728, while clinically meaningful discrepancy was highest for chloride at 31.4%.
Why this matters globally
Point-of-care testing is central to critical care worldwide. The study demonstrates why analyser-, analyte- and population-specific agreement matters more than correlation alone.
Thai researcher contribution
Prince of Songkla, Walailak, Nakhon Phanom and Mahasarakham University researchers produced a large AKI dataset addressing bedside-result safety.
Limitations to consider
This was a single-centre retrospective period spanning 11 years. Pre-analytical handling, dilution, timing and analyser-era changes may contribute. Patient harm or outcome effects were not assessed, and results are instrument- and population-specific.