Ivermectin is a worldwide-used antiparasitic drug largely administered to cattle as a topical formulation (pour-on). The actual plasma and faecal disposition of pour-on ivermectin in cattle was documented using an original pharmacokinetic model, and taking into account the oral ingestion of the topical drug following physiological licking as a secondary route of exposure. Six pairs of monozygotic twin cattle received successively one i.v. and two pour-on administrations of ivermectin at a 3-5-month interval. For one pour-on administration, the twins were separated into an unrestrained group and a group where self- and allo-licking were prevented. Ivermectin concentrations in the plasma and faeces were determined by HPLC. Licking resulted in a high intra-and inter-individual variability of systemic exposure after topical application. By the means of pharmacokinetic modelling, we showed that 58-87% of the pour-on dose was ingested, while only 10% was absorbed percutaneously. Approximately 72% of the ingested ivermectin transited directly into the faeces, resulting in a 7-fold higher faecal excretion of the parent drug than in the non-lickers. We conclude that topical administration does not guarantee a controlled drug delivery in cattle. More importantly, the simulations revealed that non-treated cattle could get easily contaminated by allo-licking, raising the public health problem of unexpected drug residues in edible tissues.