Are biters sick? Health status of tail biters in comparison to control pigs

Abstract

Background

Tail biting is a multifactorial problem. As the health status is one of the factors commonly linked to tail biting, this study focuses on the health of identified biters. 30 (obsessive) biters are compared to 30 control animals by clinical and pathological examination as well as blood and cerebrospinal fluid samples. In that way, altogether 174 variables are compared between the groups. Moreover, connections between the variables are analysed.

Results

In the clinical examination, 6 biters, but only 2 controls (P = 0.019) were noticeably agitated in the evaluation of general behaviour, while 8 controls were noticeably calmer (2 biters, P = 0.02). Biters had a lower body weight (P = 0.0007) and 13 biters had overlong bristles (4 controls, P = 0.008). In the pathological examination, 5 biters, but none of the controls had a hyperceratosis or inflammation of the pars proventricularis of the stomach (P = 0.018). However, 7 controls and only 3 biters were affected by gut inflammation (P = 0.03). In the blood sample, protein and albumin levels were below normal range for biters (protein: 51.6 g/l, albumin: 25.4 g/l), but not for controls (protein: 53.7 g/l, albumin: 27.4 g/l), (protein: P = 0.05, albumin: P = 0.02). Moreover, 14 biters, but only 8 controls had poikilocytosis (P = 0.05). Although not statistically different between groups, many animals (36/60) were affected by hypoproteinemia and hyponatremia as well as by hypokalemia (53/60) and almost all animals (58/60) had hypomagnesemia. For hypomagnesemia, significant connections with variables linked to tail damage and ear necrosis were detected (r_s/V/ρ ≥ 0.4, P ≤ 0.05).

Conclusion

The results suggest that behavioural tests might be helpful in identifying biters. Moreover, cornification and inflammation of the pars proventricularis is linked to becoming a biter. Furthermore, the results highlight the need for appropriate and adjusted nutrient and mineral supply, especially with regard to magnesium.

Introduction

It is by now generally accepted that tail biting is a multifactorial problem [1, 2]. For example, access to feeder and drinker, food and water quality, feed composition, group composition, thermal comfort, handling, access to enrichment and rooting material, space, noise, genetics and the general health status as well as certain diseases have been linked to the occurrence of tail biting [3, 4]. While many studies have been carried out, up to now, no safe prevention or cure can be guaranteed. Instead, managing the respective risk factors on a farm-to-farm basis can probably be regarded as current state of the art [5]. Moreover, different hypotheses concerning the pathogenesis have been proposed [6]. However, given the multifactorial and multi-layered nature and complexity of the behaviour, it seems likely that also different forms of pathogenesis exist. In fact, Taylor et al. [3] proposed the existence of at least three different causative behaviours of tail biting: two-stage, sudden-forceful and obsessive. While the two-stage form can be linked to the high motivation to explore and a redirection of that motivation to pen mates, the sudden-forceful form is linked to resources and the inability to reach these resources (e.g. caused by feed outages). The obsessive form is specifically linked to single animals that become a biter. This form poses the major question why single animals of the same genetics, living under the same conditions as the others become tail biters. In fact, Taylor et al. [3] highlight the need for concentrating more on the biters in scientific studies. Indeed, in the last years, a significant rise in studies concentrating on the biter can be seen: For example, Zonderland et al. [7], Beattie et al. [8], Hoy et al. [9] and Brunberg et al. [10] have focussed on behavioural differences between biters and control animals. These studies registered for example more rope directed behaviour in a tail-chew test [8], a rise also in other abnormal behaviours such as belly nosing and manipulating other parts of the body of pen mates [9, 11]. In addition, Zonderland et al. [7] registered biters to be more often in a kneeling/sitting position. Some studies have developed and used comparable ethograms to identify biters [9, 11,12,13]. All this clearly demonstrates the possibility, but also the necessity to identify and focus more on the biters especially with focus on their health status. From other species, it is well known that the health status may well be linked to the development of abnormal behaviour patterns (e.g. [14,15,16,17]). Start for treatment of behavioural problems should therefore include a thorough clinical examination to reveal potential underlying medical causes [18, 19]. As already stated, health status is commonly named as risk factor in the occurrence of tail biting [20]. For example, Fritschen and Hogg [21] hypothesised that sickness-induced discomfort may be causative for the development of tail biting behaviour. Likewise, Moinard et al. [22] proved a correlation between diseases and tail biting occurrence on farms, which also led the authors to hypothesise that sickness-induced discomfort may cause tail biting. Therefore, in this study, we aimed at identifying biters and analysing their health status in comparison to animals not affected by tail biting. Therewith, we followed the hypothesis that subclinical diseases might cause the tail biter, in particular in the obsessive form of tail biting, to start this behaviour.

Results

Results of pairwise comparisons

In the time period of the study, on farm 1, three biters were identified, on farm 2, six and on farm 3, 21. Of all 60 animals, 26 were gilts (Biters: 15, Control: 11), 26 boars (Biters: 12, Control: 14) and 8 were castrates (Biters: 3, Control: 5). The weight of the animals ranged between 9.75 and 56 kg with a mean of 22.4 (biters) and 24.2 (controls) kg, respectively. Hence, most, but not all biters were identified in the rearing period. In Table 1, the mean prevalence (mean value incl. standard deviation for continuous variables and number of affected animals for binominal variables) is presented as well as the P-value resulting from the respective statistical analyses (depending on the distribution scale of each variable). It is a comparative presentation of the Biter and Control group. Significant differences were found for the following variables: More control animals [8] than biters (2) were scored as being specifically calm in the clinical observation (variable 2), while more biters [11] than controls (5) were scored as being specifically agitated (variable 3). Overlong bristles (variable 17) were found more often in biters [13] than controls [4]. More controls [5] than biters [1] had a partial tail loss (variable 27). In the clinical examination, on average, the biters were scored slightly lighter (BCS: 2.9 (± 0.4)) than controls (BCS: 3.1 (± 0.3)) (variable 33), which could be confirmed in the pathological examination, in which the body weight of the full carcass was assessed (variable 38) and on average, biters were lighter (22.4 (± 11.7) kg) than the controls (24.6 (± 12.0) kg). The adrenal gland weight (right: variable 40, left: variable 41) was significantly higher for the controls (right: 1.67 (± 0.66) g/left: 1.90 (± 0.91) vs. right: 1.45 (± 0.60) g/left: 1.63 (± 0.70)) and also when corrected for body weight, the relative adrenal gland weight of the right adrenal gland was still significantly higher in the controls (0.00007 (± 0.00001)) compared to the biters (0.000068 (± 0.00001)). While five biters, but no control animals were affected by a hyperceratosis of the pars proventricularis (variable 90), significantly more control animals [7] than biters [3] were affected by gut inflammation (variable 91). In the blood sample, although within normal range, biters had slightly lower values for mean corpuscular volume (MCHC, variable 120; 321.8 (± 58.9) g/l) than controls (335.0 (± 9.6) g/l), were more often affected by poikilocytosis (variable 132; 14 biters, 8 controls), had lower protein (variable 135; biters: 51.6 (± 4.8) g/l, controls: 53.7 (± 6.6) g/l) and albumin (variable 136; biters: 25.4 (± 5.2) g/l, controls: 53.7 (± 4.5) g/l) levels, higher creatinine kinase (CK, variable 138) levels (biters: 418.4 (± 1082.4) U/l, controls: 310.3 (± 190.0) U/l), slightly higher glucose (variable 147, biters: 6.1 (± 2.4) mmol/l, controls: 5.5 (± 1.9) mmol/l) and slightly lower phosphorus levels (variable 152; biters: 2.5, controls: 2.6). No significant differences were found in the analysis of the cerebrospinal fluid. Although no statistically significant differences between the groups were detected, it should further be noted that many animals (53/60; 25 biters, 28 controls) had a clinical anemia (variable 158), more than half (36/60; 20 biters, 16 controls) were affected by hypoproteinemia and hyponatremia, respectively and almost all animals (58/60; 29 biters, 29 controls) had a clinical hypomagnesemia (variable 160). Moreover, 53/60 animals had a hypokalemia (variable 162). It should further be noted that although the general aim was to identify control animals that were unaffected by tail biting, in the clinical examination five controls were affected by skin irritation on the tail, six controls by bleeding of the tail (variable 25) and five controls by partial tail losses (variable 27). Moreover, two controls were scored as having necrotic changes (variable 26) and eleven with crusts (variable 28) on the tail. Likewise, in the pathological examination, 18 controls were found with dermatitis (variable 94) and five with blood (variable 96) on the tail and another six were diagnosed to have tail necrosis (variable 97) and eleven were with crusts on the tail (variable 95). In all of these variables linked to the tail of the pigs, also biters were affected.

Table 1 Mean prevalence (including standard deviation or number of affected animals) of each variable as well as achieved P-value in the respective statistical comparisons for each variable as comparative presentation for the Biter and Control group

Connection between collected variables

Meaningful and strong connections between the variables are visualised in Fig. 1 a–d. Only for the variable 83 (Lnn. gastrici: sinus histiocytosis), 84 (Lnn. sternalis: sinus histiocytosis), 112 (Cystitis), 128 (Monocytes) and 129 (Normoblasts), no connections with other variables were found at all. Apart from that, most of the connections were positive and meaningful (respective statistical parameter: ≥ 0.4, colour code: light green), but not strong (respective statistical parameter: ≥ 0.6, colour code: dark green). Only few negative meaningful and strong connections were found.

Fig. 1

Connections between variables: clinical examination (a), pathological examination (b), blood sample (c), cerebrospinal fluid (d). Connections between all variables, calculated by Spearman’s correlation coefficient (numerical variables), Chi² Test and Cramer’s V (binominal variables) and point-biserial correlation coefficient (connections between numerical and binominal variables). A connection was interpreted as meaningful if the values of the respective correlation coefficients were ≥ 0.4 (light green) and as strong if the values of the respective correlation coefficients were ≥ 0.6 (dark green) and P-values were ≤ 0.05. Negative connections are marked in transparent light and dark green colours, respectively. Due to the high number of variables, the figure is split in four parts according to the health variables under observation: a shows the variables of the clinical observation, b of the pathological examination, c of the blood tests and d of the analysis of the cerebrospinal fluid. Only those variables that had a connection to other variables are shown

Discussion

Following the differentiation by Taylor et al. [3] into (1) two-stage (linked to the high exploration motivation of the pigs), (2) sudden-forceful (linked to stress induced by uncontrollable environment features such as sudden temperature in- or decreases, feed outages, water availability etc.) and (3) obsessive (single animals bite aimfully and forcefully into tails of pen mates without a visible reason), the aim of the present study was to focus and identify obsessive tail biters. As in general, the health status is commonly linked to tail biting outbreaks, the general hypothesis was that an impairment in the health status (potentially subclinical) is a causative factor for a specific animal to start tail biting, hence, in the identified obsessive tail biters, compared to the control animals, a deviation in the health status was expected. No further specification of that general hypothesis was possible, i.e. it was not focussed on a specific organ system, but a general overview of different health parameters was provided, which lead to a large amount of variables under observation (n = 174) linked to the general health status. Therefore, this study is to be seen as explorative study analysing the general link to the health status with the further aim to be able to highlight health parameters of specific importance to be analysed in more detail in future studies.

Pairwise comparisons

In the general evaluation of behaviour during the clinical observation, significantly more biters were scored as specifically agitated, while significantly less biters were scored as specifically calm. Other studies have linked tail biting outbreaks with a generally higher level of activity and unrest a few days before a tail biting outbreak [23, 24]. Other authors have also found behavioural differences of tail biters [7,8,9,10]. These findings suggest that a standardised behavioural characterisation of tail biters, e.g. by behavioural tests for a better identification of tail biters may be possible in the future. Given the recent advances in precision livestock farming also in the pig industry [25], there is hope for a future automatic detection and early warning scheme for an animal to become a tail biter by behavioural characterisation. This is of specific importance, as the early identification and removal of biters from the group is currently probably the most effective management intervention strategy in tail biting [26]. Given the role of this study as a pilot study, the findings concerning the evaluation of general behaviour suggest that future studies should concentrate more on the behavioural differences, e.g. by proving secure identification of tail biters by standardised behavioural tests. This is further supported by the general knowledge that behaviour is one of earliest signs for changes [27, 28], the knowledge that capturing subtle behavioural changes is of utmost importance for early management intervention schemes [27] as well as the proven role of behaviour as iceberg indicator in welfare assessments in pigs [29, 30].

The weight of the adrenal glands was included as indicator, as there are hints from literature, that an enlargement can be seen as stress indicator [31, 32]. The expectation was that the tail biters were more prone to chronic stress and thus had larger adrenal glands. However, the opposite finding was made: the adrenal glands of the controls were significantly heavier. However, as also the body weight of the controls was higher, a correction for body weight was carried out. However, still, the right adrenal gland was significantly heavier in control animals. This may possibly be explained by an unsuitability of this indicator. Just looking at the size as single indicator may be a too simplified approach given the complexity of the biological endocrine systems. Probably, approaches such as the pathohistological examination with regard to the medulla:cortex ratio [33] or the gene expression in the adrenal glands [34, 35] would have been more suitable.

Significantly more biters were scored as having overlong bristles and in general, tail biters were significantly lighter than the control animals. This could be interpreted as tail biters being more often in the status of runts [36] and has already been hypothesized or reported by others [37, 38]. This may be a sign of a general shortage in nutritional requirements, such as nutrient or mineral supply or else, a health challenge in earlier stages of life. However, in the latter case, more anomalies in the pathological findings specifically of these animals would have been expected also when looking at the connections between the variables. So the more probable explanation is that for unknown reasons, nutritional requirements were not met. This comes as a surprise as all farms complied with national recommendations concerning feed composition in the different phases. However, it may be that lack of absorption occurred in these animals Kerr et al. [39] proved that pigs fed a lower protein diet showed an increase in activity level. Moreover, insufficient protein-content in feed and a generally increased activity level will lead to lesser weight gain (i.e. the animals will be the lighter ones) [40]. However, it should not stay unmentioned that despite the findings that biters were in general lighter than the controls, the general characterisation as an animal as being clinically thin (variable 39) failed to reach statistical significance in this study, which raises the question whether the small deviations in Body Condition Score and weight are really of biological significance. Moreover, although control pigs were randomly chosen that were of the same age class as the biters, this result may also be caused by the study design.

In general, one possible health challenge that might be linked to tail biting is a nutritional challenge, i.e. health issues of the gastrointestinal tract. This hypothesis is supported by the fact that tail biting in pigs often occurs around the time period two weeks after weaning [41]. Especially in pigs, weaning is carried out very abruptly and at a very early age which may well lead to severe nutritional adaptive challenges overstraining the adaptation capabilities and thus leading to behavioural disorders. However, while the finding that more biters are diagnosed with a hyperceratosis of the pars proventricularis supports this hypothesis, the fact that significantly more gut inflammations were observed in controls does not. Nevertheless, the pars proventricularis, the area of the stomach in which often gastric ulcers are observed in older pigs, should be observed in more detail also with regard to tail biting.

In the findings in the blood samples, the significant differences found for the mean corpuscular hemoglobin concentrations (variable 138; normal range: 300–350 g/l), glucose (variable 147; normal range: 4.0–6.4 mmol/l) and phosphorus (variable 153; normal range: 1.3–3.3 mmol/l) are most likely not of biological significance as they are so small and still well within the normal range in pigs. Although, the values for creatinine kinase (CK) (variable 138, normal range: 50–999 U/l) are well within the normal range and it is well known that already small influences by the sample collection [80], e.g. more flight attempts by the sample collection, may cause a rise, the difference is far more pronounced and especially the large standard deviation in biters in comparison to that of the control group should be noted. In contrast to that, more attention should be put on the protein (variable 135, normal range: 55–86 g/l) and albumin (g/l) levels in the blood sample, which were on average lower for the biters and below the reference levels (biters and controls for protein, only biters for albumin). The high need for meeting the nutritional requirements especially in fast growing breeds has already been highlighted by multiple authors [42] and moreover especially the protein and amino-acid requirements have been linked to tail biting outbreaks [4, 43]. However, clinical hypoproteinemia (variable 159), although present in more than half of all pigs (36/60) failed to reach significance in the comparison of the two groups. Hence, whether these findings are of biological significance and have explanatory power in the question, why obsessive tail biters start the behavioural disorder, must be confirmed in further studies. Another important finding from the blood samples was that significantly more biters were affected by poikilocytosis. Poikilocytosis describes an abnormality in erythrocyte forms. It has been linked to several diseases, specifically enteric diseases in pigs [44, 45]. Moreover, it has been linked to anemia caused by iron deficiency [46,47,48] and Konopel [49] has further linked it to a vitamin D deficiency. However, Christopher et al. [50] and Harvey [51] have suggested that poikilocytosis in young ruminants such as goats and cattle and maybe also pigs can be a normal finding. In the present study, the only connections of poikilocytosis was found to scratches on the ear (variable 46) and hypoproteinemia (variable 159). It should further be taken into consideration that in the present study, exactly half of affected pigs were scored as having mild poikilocytosis (≤ 33% of visible erythrocytes of abnormal shape) and the other half as moderate (≤ 66% of visible erythrocytes of abnormal shape) (results not shown), hence, the clinical relevance remains unclear.

In none of the findings, however, all biters were affected. Hence, the respective health issues cannot be the only explanation for an obsessive tail biter to be affected by the behavioural disorder.

The fact that many animals in this study (biters and controls alike) were diagnosed to have hypoproteinemia, hyponatremia and/or hypokalemia supports the hypothesis that tail biting may be linked to nutritional or mineral deficiencies. This holds despite the fact that national recommendations concerning feed composition in the different phases were followed carefully by all farms. Follow-up studies designed as controlled feeding-trials must clarify whether an adjustment of recommendations could be advisable and whether lack of absorption of nutrients could be the reason for these findings. Although no significant differences between the groups were detected, it should be born in mind that in the present study, all animals came from a population in which tail biting occurred. A special role therein plays magnesium, as almost all animals were diagnosed to have a clinical hypomagnesemia in this study. Some studies have proven beneficial effects of magnesium supplementation on different maladaptive behaviours in pigs [52,53,54]. However, another possible explanation for the high number of affected animals is that an adaptation of the reference values is needed, as usually, reference values are set from studies of limited pig population size and given ongoing advancements particularly in genetics and feeding may need regular adaptions.

Connections between variables

Most relationships between variables were only moderate. Moreover, not all variables were well connected to each other (e.g. from the clinical and pathological examination). Hence, this analysis also proves that looking at this large variety of variables was (and is) necessary. Most connections were to be expected and can well be explained by already known linkages between health parameters. However, this detailed analysis also revealed some connections that seem to be of specific importance with regard to tail biting, which will be discussed in more detail in the following:

Changes in the tail and ear linked to necrotic findings are also linked to findings linked to respiratory diseases as well as abnormalities in the skin condition, whereby this holds especially for ear necrosis as well as tail dermatitis and blood on the tail. This linkage may well be explained by a generalised necrotic occurrence as for example already described in the 1980’s by Richardson [55], Schrauwen et al. [56] and Troxler [57] and named “Swine Inflammation and Necrosis Syndrome” recently by Reiner et al. [58]. However, Reiner et al. [58] described more signs of necrosis which were not observed in the present study, although Reiner et al. [58] also states that not always all signs are observed. However, this connection may also only be caused artificially without biological meaning, as these changes are changes of rather high occurrence and further signs for a generalised necrotic syndrome were not observed in the pathologic examination. While there is a connection between findings on the tail in the clinical and pathological examination, this does not hold for all tail related variables. This is most probably due to the fact that in the pathological examination, also pathohistological findings may be included, especially given necrotic findings. Another interesting observation is that the tail and ear associated variables (variables 25–30 and 96–103) are linked to anemia, hypoproteinemia, hypomagnesemia and hypokalemia (variables 159–163), underlining the already discussed importance of nutrition and mineral requirements with regard to tail biting.

Limitations of the study

The first main limitation of this study is the small sample size. Altogether, 60 pigs were identified on three farms in Northern Germany. At the same time, on these animals, a large number of variables was assessed, which leads to the risk of α-error accumulation (for the pairwise comparisons as well as for the analysis of connections between the variables). However, given the pilot study character of this study, a larger sample size was not possible due to ethical considerations and moreover, as no information about the expected prevalences could be made beforehand. On the other hand, again given the pilot study character of this study, the large number of variables was necessary as no assumptions about expected organ systems could be made beforehand.

The second main limitation of this study is that all animals, biters as well as controls, were derived from farms in which, obviously, tail biting was present. This becomes also evident by tail damages of the control animals. Munsterhjelm et al. [59] also discussed this limitation. However, the problem is that tail biting is an unpredictable occurrence in pig husbandry. So basically, there is no farm that is 100% free from the risk of a tail biting outbreak in the future. At the same time, there is a need for studies for setting reference values and normal prevalences for the variables under study.

It must further be discussed whether there would have been even more variables of interest with regard to their connection to tail biting. For example, regarding the weight of the adrenal glands, potentially also a gene expression analysis or else a more thorough histological analysis of the cortex:medulla ratio might have yielded more insightful results. Likewise, analysing dopamine and serotonin content in the cerebrospinal fluid may be questioned as these neurotransmitters are not always freely measurable and furthermore may be dependent on the receptor density, 5-HT metabolism or gene expression in the brain [60,61,62,63].

Conclusion

The aim of the present study was to answer the question whether a diminished health status was causative in the development of the behavioural disorder tail biting in pigs, i.e. the tail biters were affected by a – potentially subclinical – disease. No obvious affection of a specific organ system could be detected. The main findings include that biters differed significantly in their behaviour as compared to control animals, in particular their general behaviour was more often described as specifically agitated and significantly less often as calm. Moreover, although pairwisely allocated, i.e. of the same age class, biters were lighter than control pigs and had more often overlong bristles. This information gives hope that in the future, an easier identification of biters will become possible. It furthermore underlines the importance of understanding and watching out for subtle behavioural changes. Moreover, biters had more often a hyperceratosis of the pars proventricularis as well as a poikilocytosis, both findings need to be studied in more detail in the future as the link to tail biting remains unclear from the present results, especially as controversely, more controls had a gut inflammation. Although no significant differences between groups were found, many animals were in a nutritional and/or mineral deficiency, which highlights the link of tail biting to nutrition and underlines the importance of exact adaption of nutrition and mineral balance.

Materials and methods

Animals and variable collection

Data collection was carried out from May 2019 to February 2021 on three conventional farrow to finish pig farms in Schleswig–Holstein, Germany. All farms kept commercial cross breds (Duroc or Piétrain × (Large White × Landrace)). All animals were fed ad libitum, however, in dependency of the farm, the unit and the phase (rearing, growing, finishing) feeding differed (but it was the same for biters and respective associated control pigs). It was either mash or dry feed ad libitum with an animal to feeding place ratio of 1:1, 2:1 or 4:1. Two farms produced their own feed while one farm bought standard commercial feed from a local provider. Feed composition was in accordance with standard national recommendations (DLG, 2021) [64] for all phases and regularly checked by the farm managers as well as the respective advisory services. Two farms routinely castrated male pigs, the other farm raised intact boars and only castrated occasionally for educational purposes. Castration procedure was carried out according to German law requirements. All pigs that were identified for the study had undocked tails. One standard management intervention scheme in the case of tail biting on all farms was the early identification of biters and removal of those animals from the group. However, for standardisation, all involved staff members were trained to use a joint ethogram for identification of biters in this study. This ethogram worked in a two-stage process: in the affected pen, at least two different pigs with bleeding tails had to be present. These pens were then observed directly for 30 min. During this time frame, any tail-in-mouth behaviour was counted as tail biting event. To be identified as biter, one animal had to bite at least four times in the tails of at least two different pen mates. Upon identification of a biter, the animal was removed from the group. The biter as well as a control animal that was randomly chosen from a pen without known signs of tail biting (no damaging behaviour had been observed/noticed by the time) but of the same age class as the biter were then transported to the University of Veterinary Medicine, Hanover, Foundation, Germany (TiHo). After a resting period of about 2 h, the animals were examined by always the same person blind to the groups. Therefore, the standard protocol of the Clinic for Swine, Small Ruminants and Forensic Medicine and Ambulatory Service (TiHo) for veterinary clinical health checks was used. On the next day, the animals were put under ketamine-azaperone-injection anaesthesia (20 mg/kg bodyweight (BW) ketamine intra muscular (i.m.), Ketamin 100 mg/ml, CP-Pharma, Burgdorf, Germany, 2 mg/kg BW azaperone i.m., Stresnil 40 mg/ml, Elanco Germany GmbH, Bad Homburg, Germany) and blood samples were collected from the V. cava cranialis as well as cerebrospinal fluid samples lumbosacrally. Directly thereafter, the animals were euthanized by administering a letal dosis pentobarbital intravenously (80 mg/kg BW (< 30 kg BW) resp. 40 mg/kg BW (> 30 kg BW) pentobarbital intravenous (i.v., V. cava cranialis), Euthadorm 500 mg/ml, CP-Pharma, Burgdorf, Germany) and subjected to a thorough pathological (including histopathology) examination, which was carried out according to the standard protocol of the Department of Pathology (TiHo). Again, the examiners were blinded to the group (biter/control). The blood samples were further analysed by the standard in-house procedure of the laboratory of the Clinic for Swine, Small Ruminants and Forensic Medicine and Ambulatory Service of the TiHo. EDTA blood was used for a large blood count and Serum/Heparin plasma for a clinical-chemical analysis. Details on the in-house blood analysis can be found in Humann-Ziehank et al. [65]. Data concerning the blood values were evaluated using the internal species-specific reference values of the laboratory. Samples of the cerebrospinal fluid were stored at -73 °C and further processed later on. The storage period of the samples did not exceed six months. For the determination of monoamines, the cerebrospinal fluid samples were deproteinized with perchloric acid before centrifugation. An aliquot of the supernatant was analyzed by high pressure liquid chromatography (HPLC) with electrochemical detection. This analysis was performed as duplicate determination; the mean of the two determinations was used for further analysis and interpretation. A more detailed description of the methodology can be found in Kanitz et al. [60].

An overview of all collected variables and a short description can be found in Table 2. In the clinical as well as in the pathological examination, in theory, other variables could have been included as well, if other findings had occurred in the animals as the protocols include the thorough examinations of all organ systems. Hence, in Table 2, only variables linked to the health status are listed that were actually observed in at least one of the animals in this study. Variables were either on a continuous scale or else, categorised to be binominal (absence, presence).

Table 2 Overview of assessed variables on the animals

Statistics

All statistical analyses were carried out using SAS® 9.4 [66]. As the experimental design was that whenever a biter was identified, a control animal was selected, pairwise comparisons were carried out. Each of the 174 variables was treated as independent variable and analysed separately. For each numerical variable, first the distribution of the differences of the pairs was tested by Shapiro–Wilk Test. In the case of normal distribution, paired t-Test was carried out, if normal distribution was not given, Wilcoxon signed rank test was used. For binominal variables, McNemar’s Test was carried out. The analysis of relationships between variables was again dependent on type of variable: The connections between two numerical variables were analysed by Spearman’s rank correlation coefficients. Connections between two binominal variables were evaluated via Chi² Test and Cramer’s V and for analysis of connections between numerical and binominal variables, point-biserial correlation coefficients were calculated. In all cases, values were evaluated as meaningful connection if the according statistical parameter reached values of ≥ 0.40 and as strong connection if values were ≥ 0.6. The level of significance was in all cases set to P ≤ 0.05.