Plasma hemoglobin (pHb) is measured when hemolysis is suspected as a result of biochemical, immunological, and mechanical conditions. In this study, we evaluated the clinical feasibility of the hemolysis index (H-index) for pHb measurement using automated chemistry analyzers. A total of 176 plasma samples were analyzed for pHb in the Severance Hospital using a Lambda 365 UV/Vis spectrophotometer (PerkinElmer, USA) and Fairbanks method 2. The remnant samples after pHb measurement were then analyzed for total bilirubin, and H-indices were determined using the automated Atellica CH930 system (Siemens, Germany) and Cobas c702 system (Roche, USA). The equivalence and correlation among the methods were evaluated using Passing-Bablok regression analysis and Spearman’s correlation analysis. All the methods showed high correlations with each other. However, the H-indices showed a negative constant bias compared to the Fairbanks method 2. When a cut-off value of 33.2 mg/dL was applied for diagnosis of hemolysis, 13 samples were determined positive by Fairbanks method 2 but negative by H-indices. These discordant results occurred mostly among samples with total bilirubin levels higher than 3 mg/dL (12 out of 13 samples). The correlation between total bilirubin and pHb measured by Fairbanks method 2 showed the largest correlation coefficient (R =0.347) and slope. These results suggest that the pHb measured by Fairbanks method 2 is more prone to bilirubin interference. In summary, our results suggest that H-indices from both automated chemistry analyzers possess excellent clinical feasibility for pHb measurements and have minimum total bilirubin interference compared to traditional measurement via Fairbanks method 2.

Keywords: Plasma hemoglobin, Hemolysis index, Spectrophotometry