Evaluation of an in-house centrifugal hematology analyzer for use in veterinary practice

Objective-To compare CBC results obtained by use of an in-house centrifugal analyzer with results of a reference method. Design-Prospective study. Sample Population-Blood samples from 147 dogs, 42 cats, and 60 horses admitted to a veterinary teaching hospital and from 24 cows in a commercial dairy herd. Procedure-Results obtained with the centrifugal analyzer were compared with results obtained with an electrical-impedance light-scatter hematology analyzer and manual differential cell counting (reference method). Results-The centrifugal analyzer yielded error messages for 50 of 273 (18%) samples. Error messages were most common for samples with values outside established reference ranges. Correlation coefficients ranged from 0.80 to 0.99 for Hct, 0.55 to 0.90 for platelet count, 0.76 to 0.95 for total WBC count, and 0.63 (cattle) to 0.82 (cats) to 0.95 (dogs and horses) for granulocyte count. Coefficients for mononuclear cell (combined lymphocyte and monocyte) counts were 0.56, 0.65, 0.68, and 0.92 for cats, horses, dogs, and cattle, respectively. Conclusions and Clinical Relevance-Results suggested that there was an excellent correlation between results of the centrifugal analyzer and results of the reference method only for Hct in feline, canine, and equine samples; WBC count in canine and equine samples; granulocyte count in canine and equine samples; and reticulocyte count in canine samples. However, an inability to identify abnormal cells, the high percentage of error messages, particularly for samples with abnormal WBC counts, and the wide confidence intervals precluded reliance on differential cell counts obtained with the centrifugal analyzer.