When ML were introduced in the market, injectable ML were highly effective against P. ovis. A single treatment with injectable ML was sufficient to kill the mites and obtain clinical cure [26,27]. Other authors later advised a second injection to get the desired efficacy [8,14,28]. Moreover, a biological model of Sarcoptes infection dynamics suggests that eradication of scabies with acaricides without ovicidal activity is only possible if two treatments are given [29].
In the present study, two or more treatments with weekly intervals were insufficient to eliminate P. ovis. Treatment failure of ML against P. ovis has been reported in sheep and cattle ([3,16,19,21]. Although treatment failure can often be attributed to non-compliance with recommended treatment and control measures [18], resistance of P. ovis populations against ML is suspected [19–21] and could be the result of decades of frequent and indiscriminate use of ML.
In this study, confounding factors that could interfere with the detection of ML resistance were eliminated as much as possible. On all farms, the experimental animals were treated according to ‘good practice’, including weighing/girth measurements to determine the dose rate, use of an injectable formulation and treatment of the whole group with correct treatment interval. Prior to treatment, crusts were removed manually and the animals were sheared, to optimise treatment efficacy. Despite these efforts, poor (or no) mite count reductions were obtained after the first treatment round and multiple treatment rounds were needed to eliminate the mites and to obtain healing of the skin.
Although ML concentrations in the skin were not determined, it seems unlikely that treatment failure was due to low ML concentration in the skin of the treated animals, given the high number of treatments with short time intervals. Lifschitz [21] showed high skin levels of ivermectin in cattle after two injections with a seven-day interval. Although BB cattle are more susceptible to psoroptic mange than other breeds, ML concentrations in the plasma and skin of BB are actually higher than in Holstein-Friesian animals [30]. Based on the above, the observed treatment failure in the present study is highly indicative for ML resistance.
The reduction in MC after the first treatment round appeared to be the best parameter to discriminate between susceptible and resistant mite populations, as MC reductions directly measure the effect of the ML on the population of mites. The number of treatment rounds and the cure rate after first treatment round did not discriminate between populations, although they are derived from the reduction in MC. CI failed to differentiate between populations until several weeks after treatment, since healing of the skin after treatment is a slow process.
To date, no threshold for mite count reductions has been defined to distinguish susceptible from resistant mite populations in vivo. We have based our threshold for resistance on existing guidelines to detect anthelmintic resistance in ruminant nematodes using a faecal egg count reduction test [25,31]. Using the thresholds described by Coles et al. [31] we identified three farms with susceptible mites, one farm was classified as ‘suspected resistance’ and 12 farms had resistant mite populations. This methodology was used because of the lack of resistance guidelines for ectoparasites in cattle. The World Association for the Advancement of Veterinary Parasitology and the European Medicines Agency do have guidelines for the efficacy of new acaricides, recommending MC as efficacy parameter and a 90% mean reduction in MC after treatment as sufficient [32,33]. Compared with these guidelines for acaricide efficacy, the thresholds we used to distinguish susceptible from resistant mite isolates may seem to be harsh. However, survival of 10% of the mites could quickly result in a relapse of psoroptic mange in the treated animals, as the mite population can double in size every six days under favourable conditions [34]. Moreover, even if the WAAVP efficacy guidelines for acaricide efficacy would have been followed, ML efficacy would only be sufficient on 5 out of 16 farms (MC reduction > 90%), which supports our conclusion that ML resistance is widespread in BB farms in northern Belgium and the south of The Netherlands. The presence of ML resistance in Belgium is no surprise, since suboptimal treatments have been applied by the majority of Belgian beef farmers for some time [18]. The emergence of ML resistance in P. ovis in beef farms in different countries (Argentina, Belgium, The Netherlands) is worrying and needs further investigation. Beef farmers should be instructed about good practices in mange control to slow down further spreading and development of resistance. Research is ongoing to identify the factors that may impede or facilitate adoption of these practices among farmers [35].
An in vitro test based on exposure of mites to different concentrations of acaricides on agar plates [22] has been used to differentiate susceptible and resistant mites [19]. An in vitro test would be useful to determine the resistance status of mite populations without the need to perform a labour-intensive mite count reduction test. However, in our hands, the in vitro test failed to distinguish the mite populations with different in vivo susceptibility from each other and in vitro test results did not correlate with in vivo efficacy. Despite large differences in MC reductions and CI reductions between farms, the in vitro efficacy showed a similar pattern in all farms and LD50 values or KD50 values did not vary with in vivo efficacy. The reason for this is unknown. To our knowledge, data about the reproducibility between laboratories are not available [19,22].