A comparison study between the results of the Siemens ADVIA 120 analyzer and the manual method for differential leukocyte counts in sheep

Background The ADVIA 120 is a widely used hematology analyzer, which has not been previously validated for determining differential leukocyte counts in sheep. Objectives We aimed to compare differential leukocyte counts on the ADVIA 120 (A-Diff) with counts obtained using the manual method (M-Diff) in sheep. Methods Ethylenediaminetetraacetic acid-anticoagulated blood samples analyzed within 4 hours of collection were used. Samples with inappropriately filled tubes, detectable clots, overtly erroneous ADVIA peroxidase cytograms, and poor-quality blood smears were excluded from the study. Two independent observers compared the results of the A-Diff with those of the M-Diff. The M-diff was performed by counting 200 leukocytes on a blood smear. Results Overall, 88 samples (44 rams and 44 ewes) were included. The correlation between the A-Diff and M-Diff was high for neutrophils (r = .873, P < .001), lymphocytes (r = .863, P < .001), and eosinophils (r = .750, P < .001), and low for monocytes (r = .212, P = .048). The Passing-Bablok regression analyses revealed constant error for eosinophils [1.17%; 95% confidence interval (CI): 0.67%, 1.55%] and proportional error for lymphocytes (0.84; 95% CI: 0.74, 0.95) and eosinophils (0.85; 95% CI: 0.74, 0.96). The Bland-Altman analyses revealed negative biases of 2.4% and 3.0% for neutrophils and lymphocytes, respectively, and positive biases of 3.2% and 0.8% for monocytes and eosinophils, respectively. The extremely low basophil percentages precluded a meaningful method comparison for this leukocyte type. Conclusions The ADVIA 120 appears to perform well for determining neutrophil, lymphocyte, and eosinophil percentages in sheep as compared with the manual method. However, blood smear examinations are recommended for the confirmation of abnormal differential leukocyte count results.