Validation of the Celltac alpha automated hematology analyzer for canine and feline blood samples
Article first published online: 31 DEC 2012
© 2012 American Society for Veterinary Clinical Pathology
Veterinary Clinical Pathology
Volume 42, Issue 1, pages 11–18, March 2013
How to Cite
McDaniel, B. J., Hirschberger, J. and Weber, K. (2013), Validation of the Celltac alpha automated hematology analyzer for canine and feline blood samples. Veterinary Clinical Pathology, 42: 11–18. doi: 10.1111/vcp.12019
- Issue published online: 4 MAR 2013
- Article first published online: 31 DEC 2012
- white blood cells
Small hematology analyzers for veterinary practices improve point-of-care diagnostic testing for companion animals. Validation of these instruments is needed to ensure the accuracy of results.
The objective was to validate the Celltac alpha hematology analyzer for feline and canine blood samples.
Canine and feline blood samples were analyzed on a Celltac alpha and a Sysmex XT-2000iV. Manual methods were used for the WBC differential count, PCV, and feline platelet (PLT) count. Flagging and precision of the new instrument were analyzed. Correlation and Bland–Altman analyses were used to compare the methods.
A total of 623 blood samples (363 canine, 260 feline) were analyzed. Within-batch precision of the Celltac showed acceptable coefficients of variation (CV) for WBC count (< 4%), PLT count (< 8%), hemoglobin (HGB) concentration (< 3%), and HCT (< 3%), while precision was poor for leukocyte subpopulations. HGB concentration and WBC count had good agreement between the methods. CV for the granulocyte (GRAN) count was 2–9% in cats and 6–29% in dogs. CV for the lymphocyte (LYM) count was 8–20% in cats and 13–51% in dogs. Negative bias and a proportional systematic error were apparent for PLT count, feline HCT, and eosinophil count. Analytical error flags and incomplete results were reported for 11.8% of canine and 25.4% of feline samples.
Leukocytosis and leukopenia were reliably detected by the Celltac alpha. The instrument provided acceptable results for total WBC count, GRAN count, HGB concentration, and HCT in canine blood samples, but PLT count was systematically overestimated. In feline blood samples, both low and high PLT counts were inaccurate and a proportional systematic error for HCT led to overestimation of this variable. Imprecision for WBC differential counts was high.