Limitations of hemolysis-based correction for potassium measurement in hemolyzed whole blood samples

This study was conducted to identify potential post-analytical corrections to adjust potassium values in whole blood hemolyzed samples using the novel GEM Premier 7000 blood gas analyzer, which features an integrated module for detecting hemolysis in whole blood. A strong linear relationship was observed between HI and percentage increase in potassium concentration in hemolyzed samples, described by the equation: [Potassium variation (%)] = 0.000857 × [HI] + 1.035279; this excellent correlation (Pearson’s r = 0.978, p < 0.001) was then applied to recalculate potassium concentrations in the hemolyzed samples. A curve-fitting based on a polynomial regression (quadratic model) was not useful to generate a significantly better correlation (r = 0.980). Bland–Altman analysis showed a minimal mean difference of 0.3 % between measured and corrected potassium values, with no statistically significant overall bias (p = 0.691). However, in 17.8 % of all samples, the corrected potassium concentration exceeded the EFLM minimum total allowable error of 7.4 % (the same rate increased to 18.9% using the best-fitting model). These results highlight the limitations of post-analytical correction methods of potassium values in hemolyzed whole blood samples.