Nematode populations and the progression of drug resistance for a variety of treatment options and management practices in sheep-rearing areas of Western Australia (WA), Victoria (VIC) and New South Wales (NSW) were simulated. The simulation was done using a computer model that predicted populations of Trichostrongylus colubriformis, Haemonchus contortus or Teladorsagia (Ostertagia) circumcincta, and the frequency of anthelmintic resistance genes.
A scoring system was used to measure the success of each option in delaying resistance to each anthelmintic and in controlling nematode populations.
In summary the best option at all sites was combining the new anthelmintic (monepantel) with a triple mixture of benzimidazole, levamisole and abamectin (BZ+LEV+ABM).
The next best option according to the model was:
in NSW, rotation at each treatment between monepantel, moxidectin and BZ+LEV+ABM;
in VIC, rotation at each treatment between monepantel and BZ+LEV+ABM;
and in WA, rotation at each treatment between monepantel (used in winter) and BZ+LEV+ABM or moxidectin (used in summer–autumn).
In WA, rapid selection for resistance occurred because of summer–autumn treatments. But, if a small percentage (1-4%) of adult sheep were left untreated then this selection could be greatly reduced.
Although relatively high resistance to benzimidazole and levamisole was deliberately assumed, BZ+LEV+ABM was still effective in controlling worms and delaying resistance.
The authors concluded that, because of cost, it may not be feasible or profitable for producers to always use the combination of all drugs. However, the second- and third-best options still considerably slowed the development of anthelmintic resistance.(Source: paper abstract).  RJ Dobson, EH Barnes, KL Tyrrell, BC Hosking, JWA Larsen, RB Besier, S Love, PF Rolfe and JN Bailey (June 2011). A multi-species model to assess the effect of refugia on worm control and anthelmintic resistance in sheep grazing systems.Aust Vet J 2011;89:200–208. The authors aimed to to develop a computer simulation model that uses daily meteorological data and farm management practices to predict populations of Trichostrongylus colubriformis, Haemonchus contortus and Teladorsagia (Ostertagia) circumcincta and the evolution of anthelmintic resistance within a sheep flock. Model predictions were compared with field observations from a breeding flock in Armidale, NSW. The model was used to see if increased refugia (leaving 1–10% of adult sheep untreated in diverse sheep-grazing systems) would delay development of anthelmintic resistance without compromising nematode control.
It was found that predicted populations of Tr. colubriformis and T. circumcincta were less than those observed in the field. This was attributed to nutritional stress experienced by the sheep during drought and not accounted for by the model.
Observed variation in faecal egg counts explained by the model (R2) for these species was 40–50%. The H. contortus populations and R2 were both low.
Leaving some sheep untreated worked best in situations where animals were already grazing or were moved onto pastures with low populations of infective larvae. In those cases, anthelmintic resistance was delayed and nematode control was maintained when 1–4% of adult stock remained untreated.
The authors concluded that, in general, the model predicted that leaving more than 4% of adults untreated did not sufficiently delay the development of anthelmintic resistance to justify the increased production risk from such a strategy.
The choice of a drug rotation strategy had an equal or larger effect on nematode control, and selection for resistance, than leaving 1–10% of adults untreated.(Source: paper abstract). Read the whole of both papers for yourself. To interpret the summary findings (above and below), you need to know what assumptions were made (eg levels of resistance) i.e. read the paper. Some comments/notes from (re-) reading these papers follow. (No, I can’t remember all the details for any length of time. Dr Dobson might). From paper 1: * (Citing Barger), when considering a worm control program, farmers will consider effectiveness, cost and ease of application, with sustainability being a lower priority. Although routine worm egg count (WEC) monitoring is advocated, only a minority do it. * Using a combination of effective, unrelated drenches is a simple, sustainable way of delaying resistance. But this is not always possible for obvious reasons. * Using monepantel in combination with or rotated with existing drugs was studied, with the resistance (R)-allele frequency deliberately being set higher than might be expected for a new drug (say 1 in 1 million). Resistance was also assumed to be co-dominant so that resistance would develop faster. * Initial R frequencies for monepantel were set to 0.001 to 0.003% for the three nematode species. ML resistance was assumed to be common (R frequency 30%; which meant MOX efficacy was still 85%+)) and BZ+LEV resistance very common (R frequency 40%). All simulations ran for 20 years. * Four options simulated: monepantel+ABM+BZ+LEV; monepantel/MOX rotation; monepantel/ABM+BZ+LEV rotation; monepantel/ABM+BZ+LEV/MOX rotation. These options + Leaving 1-10% of adults untreated was also simulated, because refugia is a key issue, especially in WA. * Broad recommendations from the model generally consistent across the three states: provide a source of refugia, use combinations, rotate monepantel with combination products. From paper 2: * Teladorsagia circumcincta (Tc) and Haemonchus contortus (Hc) can cause abomasal pH to rise, reducing establishment of incoming Hc larvae. Removing Tc and Hc infections increases establishment of Hc larvae. (The model accounted for this) (Paper 2). * In paper 2, simulations were done using the multi-species model for Merino flocks at two winter rainfall areas ( Kojonup (WA (severe hot dry summers)), and Hamilton (Vic), and two summer rainfall areas ( Glen Innes and Armidale (NSW), comparing medium and high risks for haemonchosis). * Some assumptions: BZ+LEV modelled as a single gene; initial R-allele frequency 40%. Initial R-allele frequency for ML set at 3% (emerging resistance) (the current reality may be different) and assumed to be incomplete recessive. Monepantel resistance was assumed to be co-dominant and initial R-allele frequencies were set for the three nematode species at 0.001% to 0.003% so that resistance would emerge within 10 years if monepantel was used exclusively. The initial R-allele frequency for Hc had to be increased 10-fold to 0.01% because resistance would not develop within the 20-year simulation period due to high refugia. * Four control options (ie ‘Untreated’, or exclusive use of monepantel or MOX or ABM+BZ+LEV) and four drug management control options (combinations and rotations) were simulated for each site. * Leaving 1, 4, 7 or 10 % of adults untreated (animals selected at random) was also simulated. It was assumed that leaving animals < 1 year old would be regarded as an unacceptable risk. * Death rates, nematode burdens and effectiveness at delaying resistance were scored. * According to this paper (for all sites simulated), worm control can be maintained when all young sheep and 96% or more of adult sheep are treated with a highly effective (>95% effective) product or combination of drugs. As drug resistance increases/drug efficacy declines, the % of animals required to be left untreated becomes to high and would compromise worm control. e&oe 🙂 Regards