Evaluation of three methods for measurement of hemoglobin and calculated hemoglobin parameters with the Sysmex KX-21 and ADVIA 120 in Beagle dogs and Sprague Dawley (SD) rats

• Published in: Comparative clinical pathology 2013-09, Vol.22 (5), p.929-934
• Main Authors: Yufa, Miao, Dongsheng, Pan, Yan, Huo, Bo, Li
• Format: Article
• Language: English
• Subjects: Hematology; Medicine; Medicine & Public Health; Oncology; Original Article; Pathology
• ISSN: 1618-5641; 1618-565X
• DOI: 10.1007/s00580-012-1502-x

The Sysmex KX-21 uses a novel cyanide-free colorimetric method to detect hemoglobin (HGB) concentration. This measurement has not been compared with the cyanmethemoglobin method on the ADVIA 120 which is recommended as the reference method by the International Committee for Standardization in Hematology. Additionally, ADVIA 120 uses flow cytometric method to detect cellular HGB concentration. The aim of this study was to compare the three methods of HGB analysis on the two analyzers. Fresh K 3 EDTA blood samples from 32 dogs and 27 rats were included. A complete blood count (CBC) was performed on each sample using the Sysmex KX-21 and ADVIA 120. Colorimetric and cellular HGB concentrations and all calculated variables based on HGB measurement were compared using linear regression, Passing–Bablok regression, and Bland–Altman plots, using the cyanmethemoglobin method as the reference method. In samples from both species, an excellent correlation was found between the HGB results measured with the cyanide-free method and the cyanide-based method ( r  = 0.97); however, a slight mean proportional bias of 3.8 % (dogs) and 1.9 % (rats) was observed. The correlation of flow cytometric cellular HGB concentration and extracellular total HGB concentration on the ADVIA 120 was excellent ( r  = 0.97), and no significant bias was observed. Additionally, excellent to fair agreement was evident for all calculated erythrocyte and HGB variables. Thus, the cyanide-free HGB method and flow cytometric method could be used with blood samples from dogs and rats; however, the proportional bias should be considered.