RESUMEN
An automated equine fecal egg count test, known as the Parasight System, was modified for use with small ruminants. Modifications included the introduction of a short centrifugation step in a floatation medium, an adjustment in pre-test sample filtering, and training of an image analysis-based egg counting algorithm to recognize and enumerate trichostrongylid eggs. In preliminary assessments, the modified method produced trichostrongylid egg counts comparable to manual McMaster analyses of the same samples from both ovine and caprine sources. The coefficient of determination (R2) for the linear correlation between McMaster and automated counts from these samples was 0.958, and there were no significant differences when comparing counts using feces from either sheep or goats. More extensive comparison utilized ovine samples split into three groups based on trichostrongylid egg content: Low (201-500 EPG), Medium (501-1000 EPG) and High (1001 or greater EPG). Each group contained 5 samples, each of which was used to produce individual slurries that were counted 8 times each using both McMaster and the automated method. This, again, showed no difference in accuracy between the techniques, but revealed significantly higher precision, as assessed by coefficients of variation (CoV), for the automated method for determining egg counts in the Low and Medium groups. The CoV of the McMaster method was 2.2, 2.5 and 1.3 times greater than the automated in the Low, Medium and High groups, respectively. Overall, the automated egg counting system showed good linear agreement with trichostrongylid egg counts determined with the McMaster method, and demonstrated significantly better precision. This technology reduces operator error and the results presented here illustrate its utility for determination of small ruminant trichostrongylid fecal egg counts.
