ABSTRACT
The egg hatch assay (EHA) is one of the main in vitro methods for detection of benzimidazole resistance in nematode parasites of small ruminants. However, although the EHA has been standardised at the laboratory level, the diagnostic performance of this method has not been fully characterised for field screenings. In the present work, monthly variation of benzimidazole resistance estimated by EHA was surveyed over two years in three sheep flocks and in one goat and an additional sheep flock sharing the same pastures. Resistance was measured by calculating both the effective dose of thiabendazole (TBZ) that inhibited hatching of ≥50% of parasite eggs (ED50) and the proportion (Pdd) of eggs hatching at a discriminating dose of 0.1 µg/ml TBZ. Pdd exhibited higher variability than ED50, in agreement with the higher sensitivity of Pdd to changes in resistance levels. Both resistance parameters, however, were highly correlated, and their variation was similarly related to the same factors. Resistance levels differed among sheep flocks, and the resistance level of the goat flock was higher than that measured for the sheep flock sharing the same pasture. Moreover, monthly variation of resistance in goats did not mirror that recorded in sheep. Resistance levels varied seasonally, with the highest values recorded in the spring, and they were inversely related to the number of days that samples were stored under anaerobic conditions. In addition, they were directly associated with the relative abundance of Teladorsagia spp. but inversely related to the relative abundance of Haemonchus spp. After controlling for the effects of these identified factors for variation, inter-monthly sampling variation due to unknown factors was the main source of variability, accounting for more than 60-70% of variance for both resistance parameters and yielding absolute estimation errors higher than 0.06 for ED50 or 0.2 for Pdd when resistance was estimated from a single sampling. Optimum sample size, estimated from variance components, suggested that at least 4-5 samplings would be needed to halve this absolute error, whereas additional samplings would slightly increase precision but at the cost of substantially increasing sampling effort. More research is needed to identify the main factors involved in this inter-sampling variation to standardise the implementation of EHA under field conditions.
