Comparison of 3 Leukogram Determination Methods in Avian Species: Phloxine B Stain, Blood Smear, and an Automated Analyzer

Because avian blood cells are nucleated, most automated methods used in mammalian species for total white blood cell (WBC) counts and differentials are considered inaccurate. Therefore, manual methods are routinely used in birds, although this could result in variations in methods across laboratories. The objective of this study was to evaluate and compare 3 methods of avian leukogram determination: a commercial phloxine B stain method (PB), estimation from a blood smear (EBS), and an automated analyzer (Cell Dyn 3500, [CD]). Leukograms from 23 avian blood samples were compared using these methods. All samples were evaluated once by 4 observers to assess the repeatability and precision of the manual methods (PB and EBS). Analyses with the CD method were repeated 5 times on 3 samples to evaluate repeatability. The WBC counts and differentials obtained with CD were compared to the 2 other methods by calculating intraclass correlation coefficients (ICC). Agreement between WBC counts from EBS and PB and between CD and PB was assessed with Bland-Altman plots. Results based on the CD analyzer correlated poorly with the other methods. When compared with the EBS method, ICCs ranged from 0-4.3% for heterophils, 0-12% for lymphocytes, 0-23.4% for monocytes, and were equal to 0% for eosinophils. When comparing the CD with PB, ICCs for WBC counts ranged from 85.9-91.5% among observers. High interobserver agreement was seen for the leukograms obtained with EBS (ICC = 92.9%). A good agreement was noted between EBS and PB for WBC counts (ICC = 69.5-81.3%). Bland Altman plots indicated good agreement for WBC counts between EBS and PB (slope <i>P</i> value = 0.52) and CD and PB (slope <i>P</i> value = 0.13). Although less precise than PB, EBS proved to be clinically useful and was both time and cost-efficient. The CD method does not seem adapted for avian leukocyte differentials.