HematocritRuler: determinação do hematócrito de Bothrops moojeni HOGE, 1966 baseado em smartphones

Abstract

The brazilian lancehead (Bothrops moojeni), a clinically relevant venomous species in the Brazilian Savanah, requires precise hematological methods for monitoring in captivity. As hematocrit is essential in the evaluation process, it can detect some subclinical changes. However, it is traditionally performed using the manual method and requires relatively high blood volumes for some reptile species. Given this, this study aimed to validate an alternative digital method for measuring Ht using the HematocritRuler application, evaluating its accuracy, precision, and interobserver reproducibility compared to the traditional technique. For this purpose, samples from 48 healthy adult B. moojeni specimens, comprising 24 females and 24 males, were analyzed. These specimens were sourced from the commercial breeding facility Pentapharm do Brasil Comércio e Exportação LTDA (DSM group), located in Uberlândia, Minas Gerais, which is dedicated to snake venom extraction. Blood samples (n=48) were collected from B. moojeni by puncture of the vertebral venous plexus using 3 mL syringes with 24G needles, respecting a maximum volume of 1% of body mass. The samples were stored in microtubes with EDTA K2 and kept at 4°C. Hematocrit was measured by two methods: (i) micro-hematocrit ruler: centrifugation of capillary tubes filled with 50 µL of blood and manual reading; and (ii) HematocritRuler: direct and indirect measurements of tubes with 10, 15, 25, and 50 µL. Images were captured in a standardized system with controlled lighting and analyzed by two evaluators with different levels of experience. The performance of the application was compared to the traditional method to assess its accuracy, precision, and reproducibility. The results demonstrated almost perfect agreement between the methods for 50 µL (CCC = 0.9914; total error = 4.89%) and 25 µL (CCC = 0.9864; total error = 6.88%), within the limits acceptable by ASVCP (<10%). For 15 µL, statistical adjustment reduced the total error to 6.14% (CCC = 0.9879), validating its application post-calibration. Interobserver reproducibility reached CCC = 0.9982 (50 µL), and inter-device agreement (Xiaomi/Motorola) remained CCC > 0.94 even at minimal volumes (10 µL). It is concluded that HematocritRuler is valid for Ht in volumes ≥15 µL, with accuracy comparable to the gold standard, reducing sample volume by up to 70% and expanding hematological monitoring in small animals.