Evaluation of automated methods of measuring hemoglobin and hematocrit in horses

To evaluate the accuracy of 3 automated methods of determining Hct and hemoglobin (Hb) concentration, compared with manual methods. Animals-22 clinically normal adult horses of various breeds.

A blood sample was obtained from each horse. Six dilutions (representing Hct of 0, 10, 20, 40, 60, or 70%) were prepared from each sample and analyzed, using 1 of 2 blood gas analyzers or a hemoximeter (for automated determinations) or the Wintrobe macrohematocrit and cyanmethemoglobin methods (for manual determinations). Regression analysis was used to determine mean slope relationships between Hct and Hb measurements obtained by use of manual versus automated methods. Slopes were compared, using Student's t-test.

Of the 3 automated methods examined, only 1 blood gas analyzer reported Hct and Hb values that were not significantly different from those determined by use of manual methods; however, this analyzer could not report Hb concentrations below 2.5 g/dl. The other blood gas analyzer reported values for Hct and Hb concentrations that were consistently higher than those obtained by use of manual methods at Hct < or = 20% and Hb < or = 6.6 g/dl. The hemoximeter yielded more accurate results if the Hb concentration was between 6.6 and 20 g/dl.

Although there were some limitations in measuring at low Hb concentrations, the method of determining Hb concentration and Hct with blood gas analyzer 2 was more accurate than that with blood gas analyzer 1 (Hct and Hb concentration) or the hemoximeter (Hb only).