Risk factors for tail lesions in weaner pigs

Factors influencing the risk for tail lesions in weaner pigs (Sus scrofa). by Angelika Grümpel, Joachim Krieter, Christina Veit, Sabine Dippel, 2018. Livestock science 216: 219-226.


We identified five factors influencing the risk for tail lesions in weaner pigs.•

We can recommend regression tree analysis for describing tail lesion risk factors.•

Data interpretation should include information on correlations between variables.


Tail biting is a behavioural disorder in pigs which results in tail lesions. Many factors must be considered to reduce the risk for tail biting due to the multifactorial character of this behaviour. We developed a software-based tail biting management tool called “SchwIP” for analysing farm individual risk factors for tail biting in weaner pigs. SchwIP was applied on 25 conventional farms throughout Germany who kept weaner pigs in closed barns (median 1,800 weaning places). The farms were visited up to three times between August 2016 and November 2017 and a total of 368 pens were assessed. Data regarding enrichment, pen environment, feed, water, climate, health, farm management, transport and regrouping were analysed with regression tree analysis (RT) using pen level prevalence of tail lesions (%) as the outcome variable. There were five primary influencing factors for tail lesions: docking status, stocking density, daily weight gain, suckling piglet losses and number of litters mixed during weaning. The correlation between observed and predicted prevalence of tail lesions across all pens was 0.6. Most of the factors may represent combinations of influences on a farm which agree with the multifactorial nature of the problem. Even though weight gain may also be influenced by tail biting behaviour and thus be a parallel outcome, it could be used by farmers as an indicator for initiating closer examination and intervention. The use of RT for visualising complex risk factor analyses is recommendable, though their analytical suitability for clustered data should further be evaluated.