Comparative analysis between two methods for estimative platelet count in stained blood smears from dogs

Abstract

Platelet estimation in stained blood smears is essential to confirm the automated total platelet count in veterinary hematology, even when sophisticated automated techniques are used. We investigated for the first time the adequacy of platelet smear estimation using the platelet: red blood cell ratio, considering the effects of different values of red blood cell counts and platelets in the smear monolayer area. For this study, 73 EDTA canine blood samples obtained from routine hematology were selected. All samples were submitted to an automated cellular count in a veterinary analyzer (ProCyte Dx, Idexx Laboratories), followed by traditional microscopic platelet estimation in stained blood smears in 10 high-power fields (100 ×, in immersion oil). In the same smears, the relative method of platelet estimation was conducted using the platelet:red blood cell ratio obtained in a differential cell count (in at least 1000 cells) and the automated RBC count. The agreement between the traditional and relative methods was evaluated using the Passing-Bablok regression analysis and the Bland–Altman method. Compared with the automated method, the relative method demonstrated a systematic negative bias of 100 × 10³ platelets/μL and a proportional bias. No biases were observed between the traditional method for platelet estimation and the automated method. The results confirm that for dog blood samples, platelet estimation in stained blood smears is more reliable than the relative method.