Variability of McMaster and Mini-FLOTAC techniques for equine strongyle eggs count: Novel analysis by Bland-Altman method

Abstract

Abstract Fecal egg counting techniques in horses are useful in the determination of parasitic loads and anthelmintic efficacy; and have been validated in different geographical scenarios with diverse biological samples. However, until now, there has been no consensus among laboratories, due to the use of statistical analyses with methodological and interpretative limitations, with discrepant results of accuracy and precision for McMaster (MM) and mini-FLOTAC (mF) regarding the more accurate. The study of intra-variability between MM and mF at two different fecal dilutions (5:45 and 4:56), using feces from horses naturally infected with strongyle eggs (n= 130), by duplicate readings, showed high repeatability, minimal intra-laboratorial variability, high accuracy, with low coefficient of variation: mF45[2. 92% ± 0.05], mF56[3.43% ± 0.05], MM56[3.46% ± 0.05] and MM45[3.42% ± 0.05; and high concordance correlation coefficient: mF45 0.97 (0.97 - 0.99, 95%-CI), mF56 0.97 (0.96 - 0.98, 95%-CI, MM56 0.90 (0.87 - 0.93, 95%-CI) and MM45 0.93 (0.90 - 0.95, 95%-CI). Bland-Altman inter-variance analysis (n= 1030) with a single reading, between mF56-mF45 demonstrates a high concordance correlation coefficient of 0.90 (0.87 - 0.93, 95%-CI) and low coefficient of variation (5.9% ± 0.09). On the contrary, a low concordance was recorded for MM56-mF56 and MM56-mF45 concordance correlation coefficient 0. 51 (0.49 - 0.53, 95%-CI) and CCC 0.43 (0.41 - 0.45, 95%-CI), respectively; with negative performance related to egg count and a high coefficient of variation (MM56-mF56 13.2% ± 0.20) and MM56-mF45 (CV 15.3% ± 0.24). The Bland-Altman method allows for the graphical and statistical analysis of repeatability and concordance between diagnostic methods, intuitively, with the use of accuracy and precision indicators, being easy to interpret and compare.