Numerous studies describe an association at herd or batch level between tail lesions and different lesions of the respiratory tract; pleurisy (47, 62), pneumonia (47, 58, 62), and lung abscesses (58, 62, 68) (Table 1). In contrast, van Staaveren et al. (70) did not find a batch-level association between any lesions of the respiratory tract and tail lesions.

The three studies (47, 58, 62) showing batch or herd level associations between lesions related to (enzootic) pneumonia and tail lesions suggest that generalized poor respiratory health on farm might increase the risk of pigs performing biting behavior. Studies showing that pigs with subclinical respiratory disease were more prone to bite the ears and tails support this (51). This might also explain why there is no association between lesions related to pneumonia and tail lesions at individual level (23, 47, 70) (Table 1). Indeed, Mycoplasma hyopneumoniae (M Hyo), the pathogen commonly responsible for pneumonia (75), is not spread to the lungs via the blood (76), so the pathogenesis of M Hyo—induced enzootic pneumonia is most probably unrelated to tail trauma. Nevertheless, Pandolfi et al. (58) and Valros et al. (26) found an association at animal level between severe tail lesions and pneumonia. In the study by Valros et al. (26), these cases of pneumonia were, however, almost certainly due to some other pathogen as M Hyo is eradicated in Finland.

There are studies showing batch level associations between pleurisy and tail lesions (58, 62, 68). Pleuritis is possibly linked to, for example, a high herd-level prevalence of pigs sero-positive for Actinobacillus pleuropneumoniae (APP) (75). Further, at the level of the individual animal there are strong associations between pleurisy, especially in severe forms [e.g., (70)] and lung abscesses and tail lesions, especially when tail lesions are severe (Table 1). Valros et al. (26) further showed an increase in the occurrence of pleurisy, specifically in pigs with tail lesions that had healed by the time of slaughter. A study including a pathological examination of 56 pigs [selected phenotypically as biters (16), victims of biting (16), controls in biting pen (10), and controls in non-biting pens (14)] from a farm with a tail-biting problem (31) showed that most animals in all these groups (78%) had signs of respiratory inflammations. The severity of these were, however, worse in victims of tail-biting than in the other pigs. Further, Sihvo et al. (69), showed that 9 out of 35 bitten pigs (all 9 had severely bitten tails) had chronic purulent or necrotizing interstitial or bronchopneumonia with (n = 7) or without (n = 2) abscesses. At least part of these lung lesions were caused by secondary, environmental bacteria, which indicates a possible spread via the venous system from the tail. Other types of lung lesions, such as pulmonary actinobacillosis, and mild lymphocytic interstitial pneumonia were evenly spread between control and bitten animals.

There is a well-established association between tail lesions, trimming and condemnation of carcasses (partial or entire) particularly for abscessation at individual pig level (26, 30, 65) (Table 1). Indeed, the lack of an association between carcass condemnation for abscessation and severe tail lesions at individual level in the study by Harley et al. (64) is surprising given that abscessation is the primary reason for carcass condemnation in the Republic of Ireland (24). However, there was an association between mild tail lesions and partial carcass condemnation for abscessation. It is possible that the tail lesions were severe at some stage on-farm and thereby contributed to secondary infection and associated partial carcass condemnation but healed such that they scored mild at the time of slaughter. Infections of the joints (arthritis) and bones (osteomyelitis) were the second most common association between tail lesions and reasons for carcass condemnation/trimming (Table 1). The interrelationship between (poly)arthritis, vertebral osteomyelitis and tail lesions is well-documented [Meijer et al. (67) cited by (30, 63)]. Hence, it is unsurprising that most of these associations were at individual level (23, 25, 60, 66).

Farm and abattoir based studies frequently report that male pigs are more likely to be the recipients of tail-biting (i.e., to have tail lesions) than female pigs irrespective of whether they are castrated (5, 23, 25) or entire (24, 64, 92–94). However, Sinisalo et al. (22) found no difference between the sexes. There is no evidence of a propensity of a particular sex to tail bite. Nevertheless, there was a faster increase in piglets with tail damage in all-female groups compared to piglets in all-male or mixed sex groups post-weaning (94). Furthermore, the duration of tail damage was higher in males in mixed sex groups. As mixed sex groups are the norm in practice, this could explain why more males show tail damage.

The evidence surrounding the propensity for a particular sex to succumb to disease is equally tentative. Male piglets are certainly at higher risk of dying than female piglets (95). They are also more likely to succumb to diseases such as post-weaning multisystemic wasting syndrome (96) and Lundeheim (97) reported that castrates had a higher burden of respiratory diseases than gilts. There is some evidence to suggest pathogenic mechanisms through which males may be more susceptible to disease. Castrated male piglets are more susceptible than females to physiological stress (98, 99) and are more fearful (98, 100, 101) which may increase susceptibility to disease (102). Furthermore, entire male pigs show higher cytokine levels in response to a lipopolysaccharide (LPS) challenge compared to females (103).

There are a limited number of reports on breed differences in exhibition of DB, [e.g., (104, 105)] as well as in predisposition to be the recipient of tail-biting (5, 22, 33). Generally, it is difficult to separate effects of breed from effects of selection for production traits such as leanness on the propensity to bite or be bitten. Breuer (106) reported that predisposition to tail bite had a heritable component within Landrace pigs and that this predisposition has a genetic correlation with lean tissue growth rate and lack of subcutaneous fatness. Similarly, Ursinus et al. (10) found that gilts with high levels of tail and/or ear biting had a higher genetic potential for several production traits. On the other hand, there is clear evidence that genetic selection for high lean tissue growth rate is associated with changed immune response and increased disease susceptibility (81, 107, 108). Selection for high lean tissue growth rate is also unfavorably associated with leg weakness score linked to osteochondrosis in pigs (109). Finally, Hessing et al. (110) found a strongly significant “litter-effect” on gastric ulceration which might indicate a genetic predisposition.

Tail-biting pigs are often smaller “runt” individuals (23, 111, 112). Munsterhjelm et al. (113) found that performers of “tail in mouth” behavior were smaller at birth. However, there is limited evidence that low birthweight (BW) is a risk factor for DB. Ursinus et al. (10) found no difference in BW between high-tail biters (which also performed more DB directed at other body parts), medium tail-biters and non-tail biters. Similarly, a recent, on-farm study (114) failed to link BW to pig-directed manipulation in the grower/finisher periods. There is some evidence of a predisposing effect of undernutrition during lactation, and therefore low weaning weight, on subsequent manipulatory behavior of weaned piglets whereby the later tail-biting behavior seems related more to growth rate immediately preceding onset of the problem [reviewed by (80)]. This review indicated that DB may be increased in piglets originating from large litters, which would tend to have lower average individual BW, although this is more likely explained by the fact that social competition is greater in large litters. Indeed, Ursinus et al. (10) found that gilts showing high levels of tail-biting originated from larger litters. When it comes to the risk of becoming a victim of biting, Hakansson and Houe (115) recently showed that the risk for tail damage was higher in piglets with a high weaning weight.

There is strong evidence that low BW piglets, more prevalent in large litters, are more susceptible to disease in both the short and longer term. Initially there are long-term dysfunctions in vital organs in low body weight piglets [e.g., gastrointestinal tract—(116)], but they are also less competitive at the udder and ingest less colostrum (117). Piglets with low immunoglobulin G serum concentration have a low survival rate (118), although this might also reflect nutritional and thermal challenges to neonates with low colostrum intake. Calderon Diaz et al. (92) reported that pigs with a low BW (<0.95 kg) were at higher mortality risk throughout the production cycle and were at higher risk of locomotor diseases. In line with this, pigs in large litters had a slightly increased risk of developing foot abscesses, compared with pigs in small litters (119). Also Feldpausch et al. (120) reported a highly increased mortality rate in piglets of low BW. Following on from this, pigs of low BW in larger litters are therefore likely to be at greater risk of death and to be lame prior to slaughter. Furthermore, low BW is a risk factor for respiratory disease (121) and the development of PMWS (96, 122) with the latter study suggesting a threshold <1.3 kg. Calderón Díaz et al. (92) presented a strong positive correlation between birth BW and weaning BW. They found that pigs with a low weaning weight were more likely to have an increased risk of lameness, pleurisy, pericarditis and heart condemnations at slaughter. This is in line with the predisposing effect of undernutrition during lactation on subsequent manipulatory behavior of weaned piglets [reviewed by Prunier et al. (80)], suggesting a common early life background to both health and behavior problems.

Clearly, the occurrence of DB in a group negatively affects the growth of the victims. Several studies show lower average daily gain in victims of tail-biting compared to non-victims (22, 61). Similarly, Camerlink et al. (123) reported a poorer average daily gain in pigs that received a lot of oral manipulation (>2% of time), including tail and ear biting and other types of DB. On the other hand, Hakansson and Houe (115), as well as (114) found a greater risk of tail lesions in pigs with a high growth rate before and after weaning. In this case, however, the lesions were not severe, and the weight gain was recorded mainly before the lesions, which implies that high weight gain, or “nutritional need,” predisposes pigs to becoming victims of DB, instead of oral manipulation or tail damage reducing growth rate, as previously suggested. However, it is more likely that the two are not mutually exclusive; while faster growing pigs may be more likely to become victims—they may then have a reduced growth rate if severely bitten. The relationship between growth rate and risk of disease is clearer. Almost all of the common health problems/diseases of pigs are associated with poor growth [e.g., (97)]. A Brazilian study reported an association between lung lesions (pleuritis and embolic pneumonia) and locomotor problems in low body weight pigs at slaughter (59). Similarly, Kuchling et al. (124) reported that expected daily gains reduced significantly for pigs with at least one of the post-mortem findings arthritis, abscess, severe pneumonia, visceralis pleuritis and hepatitis.

The predisposing effect of undernutrition during lactation (80) corresponds to studies showing that pigs exhibiting tail-biting behavior were those that experienced a growth check (111, 112, 125). However, fast growth rates in older pigs may also play a role in the development of DB. At group level, Diana et al. (126) found that pigs with antibiotics in their diet (associated with higher growth rates), were more likely to have tail lesions and performed more ear biting than pigs without. In another study, the fastest growing pens of pigs on a commercial farm were more likely to have ear and tail lesions compared to compromised pigs (127). At individual level, Ursinus et al. (10) reported that gilts showing most biting behavior were also the largest and fastest growing animals. Similarly, Valros et al. (114) reported a high growth rate linked to a higher level of pig-directed manipulation. While pigs that are thriving and growing quickly might be more likely to engage in DB they are almost always less likely to experience health challenges (128). The only condition this does not hold true for is locomotor disorders where selection for fast growth rates is a risk factor for osteochondrosis (109). Furthermore, Quinn et al. (129) reported that gilts fed a diet formulated for slower growth rates (and resulting in reduced average daily gain and lower body weights) had improved locomotory ability and less severe humeral condyle joint lesions compared to gilts on diets designed for fast lean deposition.

Prunier et al. (80) defines personality and coping style and elucidates how they may influence the propensity of pigs to perform DB. The effect of coping style appears to be mediated by the degree of environmental enrichment (130, 131), with pigs of a more reactive coping style showing greater frustration (132) more pessimism (133) and more DB with a lack of or loss of, rooting materials. Although, Ursinus et al. (42), found no such link between coping style and DB. Fearfulness is another personality trait relating to tail-biting. Ursinus et al. (134) reported that tail biters spent less time near a bucket introduced in a novel environment, and suggested that tail biters were more fearful. Zupan et al. (34) found the opposite, with tail biters being faster to touch a novel object, lying more and walking less in a novel environment. Linking tail biting to behavior in fear tests is complex as fear-related behaviors in such tests might reflect temporary states rather than stable personality traits. Moreover, tail biting is not consistent, except for those biters that display high levels of tail biting (10, 42). This could relate to the different types of biting identified, with possibly different underlying motivations and risk factors (33).

Pigs show genetic variation in coping style associated with neurotransmission (135). Tail biters had a lower store of serotonin in blood, and a higher blood platelet serotonin velocity. However, this only held for the period in which they were classified as tail biters (134). Valros et al. (136) reported increased serotonin metabolism in the pre-frontal cortex in tail biters and, only in tail biters, a link between central serotonin and tryptophan levels in the blood. However, it is unknown if these differences in serotonin metabolism are temporary states or stable differences between pigs prone to tail bite and those with a lower propensity to tail bite.

Little is known about the link between health and personality in pigs, except that reactive pigs are more susceptible to develop gastric lesions in barren housing than proactive pigs [(137), but see (138), who found no effect] or when housed with other reactive pigs only (139), whereas proactive pigs showed more heart deviations postmortem (139). Several studies point to a difference between proactive and reactive pigs in parameters related to innate and specific immunity [e.g., (101, 140–145)], although the consequences of these differences for disease susceptibility are unknown. According to Kanitz et al. (142), when exposed to stressful situations, the effects of coping style on humoral immunity differs depending on the specific function of the immunoglobulin classes, as proactive pigs may have an increase in IgA concentration but not in IgM. This makes coping style-related differences in disease susceptibility probable.

There are numerous animal-based risk factors shared between tail-biting and disease but several of these are tentative particularly those relating to coping style and other personality traits (Table 2). Many studies show that male pigs are more likely to be bitten and there is evidence that they are more susceptible to stress (98, 99), disease [PMWS—(96)] and death (95). Furthermore, the limited evidence presented above suggests a propensity of females to bite tails. Hence, there might be sex-based differences in the likelihood of a pig to become a biter or a victim of DB, and to succumb to disease but there is surprisingly little research in either area. Birth weight is a major risk factor for disease (96, 119, 122). However, its effect on tail biting is confounded with litter size whereby piglets from larger litters, being smaller, may be more disposed to tail bite (80). This might suggest that better health and growth rate during lactation reduces the risk of a pig becoming a tail-biter in later life. However, genotypes with generally faster growth rates and lower fatness are more likely to tail bite (10, 106) and groups of pigs with high growth rates appear to be more susceptible to DB [e.g., (10, 127)]. These findings appear to contradict those suggesting that pigs experiencing a growth check are the tail-biting culprits. However, pigs in both cases may have a high “nutritional need/metabolic demand,” either because of their (history of) poor health, poor growth and undernutrition or, conversely, due to a high growth/production level demanding greater nutrients/energy [as per (80)]. This increases the pigs' motivation and competitiveness for access to feed, which in turn can lead to DB (33). Hence, it appears that rather than growth rate per se, a high “nutritional need/metabolic demand” is likely the common underlying factor linking health and DB. We elucidate this potential mechanism for a causal link below in section on Elucidating causal mechanisms.

The effects of temperature on tail biting are contradictory. Smulders et al. (6) found that a higher temperature in the nursery pens was associated with a higher number of pens containing at least one pig with a tail or ear lesion. Similarly, Holling et al. (146) found that a temporary ventilation failure resulting in an increase in temperature was associated with tail-biting. This would suggest that tail biting is higher in the summer. Indeed, Sällvik and Walberg (147) reported that high summer temperatures inside the housing tended to increase tail-biting. However, several epidemiological studies report a higher likelihood of tail lesions in the winter (60, 72, 148). It is likely that fluctuations or extremes of temperature in either direction are risk factors for tail biting and other DB.

Heat stress caused by high temperatures is associated with numerous detrimental physiological changes in the pigs body (149). However, while it undoubtedly leads to poor welfare and poor growth and reproductive performance (150) high temperatures are not directly associated with specific diseases. In this case, there are more implications of cold temperatures. There is an increased incidence of diarrhea in weaned piglets maintained at chronic, moderate cold temperatures (18–20°C) [reports reviewed by (151)]. Armstrong and Cline (152) reported that artificial infection with enterotoxigenic Escherichia coli resulted in higher incidence of diarrhea in newly weaned piglets when exposed to severe cold stress (4–10°C). Temperature fluctuation is also a risk factor—Le Dividich (153) observed a greater incidence of post-weaning diarrhea in piglets kept in a continuous (hourly) fluctuating temperature, 23.5 ± 3°C compared with a constant environment, 23.5 ± 0.5°C. Burrough (154) also identified low temperature as predisposing the development of swine dysentery in endemically infected farms. Many reports of temperature as a risk factor for disease relate to respiratory diseases and Stark (155) cites numerous examples from cross-sectional and longitudinal studies. For example, Geers et al. (156) observed a negative relationship between coughing and air temperature, while the number of treatments for pneumonia was higher for low and highly variable temperatures (157). In a more recent study (75), a mean temperature in the finishing room below 23°C was a risk factor for pleuritis. Similarly, neonatal pigs maintained in a cold environment of 18°C and administered LPS experienced a period of hypothermia coupled with altered endocrine and proinflammatory cytokine responses (158). Additionally, growing pigs kept at high ambient temperature have greater production of heat shock proteins, which are protective during inflammation (159), than those kept at thermoneutrality (160).

Whilst linked to temperature, since ventilation rate affects both, several aspects of air quality are also important risk factors for DB and disease. Various authors describe ammonia as the primary noxious gas able to induce stress and consequent DB, including tail-biting (161, 162). Scollo et al. (88) found that high ammonia was the variable showing greatest importance in influencing prevalence of tail lesions with a threshold level of 28 ppm. Ammonia is also a risk factor for liquefaction necrosis (163) and low body weight (164) most likely due to less feeding behavior (165). These authors reported elevated monocyte, lymphocyte and neutrophil counts in weaner pigs exposed to a concentration of 35 ppm using gas exposure chambers, while pigs exposed to 50 ppm showed increased levels of the acute phase protein, haptoglobin. Exposure to ammonia also affects pig activity (164). Done et al. (166) found small pathological changes in the respiratory tract of pigs exposed to concentrations of ammonia representative of commercial conditions (0.6–37.0 ppm). Meanwhile, Stark (155) cited numerous studies implicating aspects of air quality including humidity, gases, bioaerosols and dust. For example, an air exchange rate of >60 m³ per h per pig, which would reduce all detrimental air contaminants, had a protective effect on pneumonia (63). Relative humidity (RH) correlated with the number of necessary treatments for pneumonia, but with a lag of 2 weeks (157). High humidity was also a risk factor for ear necrosis syndrome (167). A high mean CO₂ concentration (>1,600 ppm) in the finishing room was a risk factor for pneumonia (82). This is consistent with previous findings, indicating that high CO₂ concentrations may have detrimental effects on respiratory tract health (168–171). Furthermore, air quality (a composite score of airflow patterns, NH₃, CO₂, RH) is also a risk factor for post-weaning digestive disorders (172).

Only, Sällvik and Walberg (147) demonstrated that cold air at high airspeed (i.e., draught) tended to increase tail-biting. Exposure to draughts also resulted in more redirected explorative behavior in weaned piglets, including ear biting, and agonistic behavior, especially head knocks with biting (173). In contrast there is good evidence that draughts are a risk factor for disease with increases in coughing, sneezing and diarrhea reported in weaned piglets (174). The setup of the ventilation system seems to be important for respiratory disease. A direct fresh air inlet from outside or from the corridor into the room vs. an indirect air inlet was a risk factor for pneumonia in weaned pigs (75) or for being seropositive to Actinobacillus pleuropneumoniae serovar 3 in finishing pigs (175). Fablet et al. (75) also found that the range of temperatures controlling ventilation rate, which affects how quickly fan speed increases and decreases and may create air draughts at the pig level, was associated with prevalence of pleuritis.

The evidence for temperature as a shared risk factor for DB and disease is conflicting. While higher temperatures are risk factors for DB (147), colder temperatures are more frequently cited in relation to disease (Table 2), particularly respiratory disease [e.g., (75)]. Nevertheless, these findings support the well-established requirement to maintain pigs in their thermoneutral zone to optimize welfare, health and performance as recently demonstrated by Chantziaras et al. (176). There is compelling evidence supporting the role of draughts and poor air quality, particularly high ammonia levels as risks for disease. However, it is only in the case of ammonia that this is shared with DB. This is in spite of strong anecdotal and farmer reported evidence of the detrimental effects of draughts (or “cold air pockets”) on tail biting (45).

Certain dietary ingredients increase the risk of tail-biting. These include whey (87, 180) and wheat (87). Interestingly, a high level of wheat in the diet is also a risk factor for gastric ulcers (182, 183) and non-specific colitis (184, 185). Two old studies report on “whey disease syndrome,” characterized by sudden death associated with acute and extensive hemorrhage into the lumen of the small intestine (186, 187). A more recent study also describes deaths caused by “haemorrhagic bowel syndrome” in pigs fed whey (188). However, this is related to poor feed hygiene practices more than whey feeding per se.

The evidence around dietary crude fiber and tail-biting focuses on undocked pigs and is dominated by strong batch and age effects (189, 190). The pigs in these studies were young and therefore generally fed ad-lib and so unlike dry sows (and possibly older pigs—see below) do not experience extreme frustration of feeding/foraging motivation. This is supported by van der Peet-Schwering et al. (191), who found no effect of replacing part of the wheat in the diet with 12–14% of fiber rich feed ingredients on indicators of tail-biting in the nursery. However, they found a lower percentage of grower-finisher pigs with tail damage and with reduced tail length, and a lower number of animals removed to the hospital area on this dietary treatment. It is not surprising that minor increases in dietary fiber level (from 4.2 to 6.6% crude fiber) with minimal enrichment provision did not control tail-biting in undocked finisher pigs on slatted floors (192). Some studies indicate a direct link between ingested fiber and gastric ulcers, e.g., (193). This positive effect is possibly mediated by increased production of saliva (194).

A widely studied aspect of feed quality is contamination with mycotoxins (195–197). There are anecdotal reports of contamination with mycotoxins being associated with tail-biting [e.g., (198)]. High levels of deoxynivalenol were found in feed and sow blood samples in herds where piglets were affected by tail necrosis (199). Similarly, both tail and ear necrosis was observed after a combination of ergotamine, ergocristine, and ergonovine (10 mg/kg of grain) was fed to older pigs (200). Such necrotic lesions may predispose pigs to perform DB but there is no published evidence of a direct link. In contrast, the immunosuppressive effects of mycotoxins and associated implications for disease in pigs is widely studied (201). Specifically, ingestion of mycotoxin-contaminated feed increases the susceptibility to infectious diseases, reactivates chronic infection and decreases vaccine efficacy (202–205). Aflatoxin B1 (AFB1) lowered the incubation period for swine dysentery and prolonged the clinical diarrhea compared to a control group (206). FB1-exposed piglets showed longer shedding of enterotoxigenic Escherichia coli (ETEC) following infection and a lower induction of antigen-specific immune response after oral immunization (207). Similarly, Pósa et al. (208) showed that FB1 aggravated the progression of respiratory disease.

The feed delivery system is often confounded with the level, frequency and form of feeding, making it difficult to clearly determine the exact nature of risks. Housing systems with ad lib feeding and multiple feed spaces have less tail-biting compared to restricted and single space feeders (6, 14, 209–212). Similarly, other factors which result in increased feed competition such as dysfunction of the feeding system (15, 90) or deficiency in timeliness of feeding (88) are risk factors for tail-biting.

Stärk (155) identified an adverse influence of floor feeding on respiratory disease (213–215). As mentioned, beneficial effects of restricted feeding for leg health are mediated by slower growth rates [e.g., (129)]. Some studies indicate that strategically restricting feed to piglets in the first few days post-weaning reduces post-weaning diarrhea [e.g., (216)] though others contradict this [e.g., (217)]. Further, the likelihood of oesophago-gastric ulcers at slaughter was higher in pigs from herds with ad libitum feeding and herds with automatic feeding (218).

Liquid feeding may predispose to tail-biting (14, 17, 87, 219). However, Hunter et al. (209) found that meal or liquid feeding reduced the probability of long-tailed pigs being tail bitten compared to pelleted feed, and Smulders et al. (6) found a higher number of pens with pigs having tail and ear lesions when pigs were fed a dry diet in the growing unit. Differences are possibly linked to differences in feeding frequency, since both Temple et al. (17) and Kallio et al. (87) identified a higher number of smaller meals per day as a risk factor for tail-biting. In agreement, Hessel et al. (220) found that a higher frequency of daily liquid feeds fed to pigs restrictedly resulted in more competition at feeding than a lower feeding frequency.

Use of pelleted compound diets vs. meal was a risk factor for tail-biting in some studies (14, 87, 209). In this respect, it is interesting that pelleting of diets poses a risk for gastric ulcers (183, 221–223) and for non-specific colitis (184, 224, 225). Dry feeding, as opposed to liquid feeding, is a risk factor for Salmonella in pigs (222). It is also generates dust which might increase risk of respiratory diseases (226).

In spite of the obvious importance of water for pig welfare, few studies demonstrate a role in tail-biting. Taylor et al. (15) highlighted a reduced risk of tail-biting when pigs had “good water access” (low ratio of pigs: drinkers, clean drinkers and good water pressure). Further, pig producers ranked “water available to all pigs” as the most important preventive factor for tail-biting (91) or ranked it very high (4th of 20 preventive measures) (46). Inadequate access to water is also a risk factor for respiratory disease (155). Madsen and Kristensen (227) revealed alterations in the circadian rhythm of the drinking behavior of young pigs that is an early indication of health impairment. Furthermore, gastric ulceration was associated with the source of water, with use of bore-hole water being a risk factor (218). This may relate to possible effects of pH, buffering capacity or microbiological quality. Urolithiasis (“stones” in the urinary system) in finishing pigs was partially associated with inadequate provision of drinking water (228).

The interrelations between all feeding related risk factors complicates extrapolation of commonality between DB and disease (Table 2). For example, feeding frequency is interrelated with feed form. Hence, while Taylor et al. (15) pointed out that an optimal number of meals/day can minimize tail-biting, the optimal number of meals would need to be determined for all feeding systems and diet forms. The inconsistency is reflected in the effects of liquid feeding, which is a risk factor for DB while dry feeding seems to be a risk factor for certain health problems. Dietary roughage (i.e., fiber from straw or silage) has numerous benefits in terms of facilitating foraging and exploratory behavior, as well as chewing which stimulates saliva production. It also has structural properties that are protective at gut level (193). Not surprisingly there are benefits both to pig behavior, e.g., (229), and health, e.g., (194). However, studies evaluating different levels of dietary fiber show less compelling benefits for tail biting or pig health possibly because growing pigs are generally fed to appetite. The detrimental effects of wheat and whey in the diet and diets in pelleted form are consistent for both DB and disease albeit not widely studied for the former. The same holds for the protective effects of non-competitive feeding systems and easy access to water of high quality. Mycotoxins are the major aspect of diet quality potentially posing risks to pig health and DB but while negative implications for immune function, if not disease itself, are numerous, the evidence for the latter is surprisingly scant.

Slatted floors are a risk factor for tail-biting (14) with both the presence and the area of slatted flooring implicated (87, 230). This might relate to effects on thermal or physical comfort, air quality or the reduced likelihood of providing bedding. Slatted flooring is also a risk factor for gastric ulcers (48, 221). This could be confounded with the absence of straw bedding, although this is not always the case (231). Stärk (155) identified reports in which slatted floors and poor floor insulation (no bedding) had a negative impact on respiratory health. Sanchez Vasquez et al. (232) demonstrated that part slatted floors compared to solid floors with bedding were a risk factor for enzootic pneumonia-like lesions and pleurisy in the carcass. Aspects of flooring, are also associated with locomotory disorders in finisher pigs (233).

Straw bedding, as well as offering thermal, physical and nutritional benefits, is also a source of enrichment allowing pigs to express foraging and exploratory behaviors. Hence, the absence of this manipulable material is a consistent risk factor for tail-biting (14, 15, 48, 87, 112, 125, 209). Even moderate bedding decreases tail lesions in undocked finisher pigs (234). Smaller amounts of manipulable substrates such as chopped straw (235), light chopped straw and wood shavings (234) or fresh barley straw (236), mushroom compost (237), peat and sawdust (238), grass silage (194), whole-crop barley and pea silage (239), hessian fabric (10) or freshly cut wood (85) also reduces penmate-manipulations/tail and ear biting. Buijs and Muns (240) summarize the non-straw enrichment material benefits for tail-biting in more detail. It is also worth mentioning that a change in the quality of enrichment between the production stages is a risk factor for tail-biting (15).

There are some concerns for pig health associated with straw provided as bedding or enrichment due to the potential presence of mycotoxins (241). Pigs are susceptible to several types of mycotoxins which have a detrimental impact on immunocompetance (201, 242) if not pig health. Other authors show health challenges for pigs on deep bedding associated with poorer hygiene, including PMWS susceptibility and respiratory/enteric health though there were positive effects on leg health (48). Provision of straw or other fibrous material also reduces the risk of gastric ulceration (194, 221, 243–246) and other stomach and intestinal disorders (247). This could be due to the beneficial effect of fiber ingestion per se (193, 246). However, Gottardo et al. (231) found the provision of other forms of environmental enrichment (hanging chains, plastic objects) also decreased the risk of gastric ulceration on farms that did not provide bedding. There is also evidence of a protective effect of environmental enrichment against other disease challenges. For example, pigs reared under enriched conditions exhibited fewer days of diarrhea after weaning (248) while those raised with a combination of social and environmental enrichment factors had a faster viral clearance and developed fewer and less severe lung lesions after an artificial disease challenge (2).

Larger farm size is a consistent risk for disease and DB (14, 83, 87, 88, 249) with Moinard et al. (14) and Scollo et al. (88) also identifying more pigs per stockperson as a risk for tail-biting.

Stärk (155) cites numerous studies in which herd size was a risk for respiratory disease, with subsequent studies also highlighting a specific risk for pleurisy (250, 251). Ramis et al. (252) suggested that an increase in gastric ulceration in pigs from large farms might be due to increased infection pressure from other diseases, but there are implications of many other non-infectious factors associated with herd size. Goldberg et al. (253) see large herd size as a significant risk factor for mortality in sows, also supported by Bergman et al. (254). Further, Munsterhjelm et al. (177) found an association between farm size and “found dead” mortality in finishers.

In epidemiological studies both Smulders et al. (6) and Pandolfi et al. (58) reported links between poor biosecurity practices and DB/tail lesions. For example, the former authors found that the absence of footbaths was a risk factor for tail and ear biting. Furthermore, tail-docked pigs subjected to low sanitary conditions showed increased ear-biting behavior and damage to ears in comparison to unchallenged animals, although this effect was diet dependent (43).

Good hygiene and health management can prevent, as well as interrupt, the spread of infection. Generally, poor biosecurity practices pose a greater risk of disease [e.g., (58, 255)]. Pen hygiene is particularly important for enteric disease. Madec et al. (172) showed that pen hygiene status on arrival posed a risk for post-weaning digestive disorders. Non-hygienic husbandry, inadequate quarantine and biosafety measures are also risk factors for PMWS (256–258). Pen hygiene predisposes development of swine dysentery in endemically infected farms (154). A Polish study reported that farms with an all-in/all-out hygiene policy (AIAO) had a significantly lower prevalence of Lawsonia intracellularis (259). In addition, the odds of having a Salmonella-positive sample was 3.9 times lower in farms practicing AIAO (222). The importance of biosecurity, including animal purchasing policy, AIAO management and hygiene in the occurrence of respiratory disease are well-known (253, 260, 261) [see also (155) for review]. Finally, pen hygiene influences prevalence of locomotory disease through increased risk of infectious agents, but also softening of the claw in wet conditions (262) and increased risk of slipping and injury on soiled floors (263).

As stocking density is the combination of pen size and number of pigs per pen, either factor separately or in combination can influence its effect. Most studies, both epidemiological and experimental, report more tail-biting in pens with higher stocking densities (14, 88, 177, 219). For example, weaner pigs with lower stocking density (<38 kg/m²) had a lower prevalence of tail lesions than pigs in pens with a stocking density ≥38 kg/m² (84).

High stocking density also increases the risk of many health problems, including non-specific colitis (224), swine dysentery (154) and respiratory disease (155), clinical leg weakness and claw disorders (264). Tuovinen et al. (265) estimated an odds ratio of partial carcass condemnation of 4.2 for a decrease of the total pen area per pig by 0.1 m². Stocking density influences not just floor space, but also three-dimensional space. More than 3.5 m³ of air volume per pig was preventive for pleurisy (63).

Group size, although sometimes confounded with stocking density, can influence disease transmission and the scope for DB. Holmgren and Lundeheim (180) found that an increase of one pig to the group increased the prevalence of tail-biting by +0.2% with long-tailed pigs, and suggested that this was a consequence of an increase in the number of potential victim pigs. Kallio et al. (87) also found more tail-biting when finishing group size was >9 pigs. Reducing the size of the group from 15 to 12 finisher pigs reduced tail-biting despite a tendency for more tail-directed behavior (266, 267).

Madec et al. (172) showed that larger group sizes (>23 pigs/pen) increased the risk for post-weaning digestive disorders, whilst the risk of PMWSwas greater with large pens (258) or more pigs per pen (122). Higher group size is also linked to risk of non-specific colitis (268) and to respiratory disorders (63, 168). Furthermore, pigs housed in large groups had poorer leg health compared with pigs housed in small groups (262, 269).

Friend et al. (270) found more tail lesions when more litters were regrouped. In agreement, Gruempel et al. (84) reported a lower prevalence of tail lesions in pens where there was less regrouping at weaning. Interestingly, Arey (271) reported an outbreak of tail-biting in pigs after regrouping in new pens with twice as much space as before. Regrouping unfamiliar pigs is a cause of stress and de Groot (272) suggested that this was responsible for suppressing the immune response to a viral vaccine. Regrouping is a risk factor for PMWS (256, 258) and for locomotory disorders (273). It is also a possible risk factor for gastric ulceration (110, 231) while more frequent movement of animals is a risk factor for respiratory disease (213, 214, 274).

Moinard et al. (14) found a higher incidence of tail-biting in farms where cross- fostering was practiced compared to farms with no cross-fostering. However, it was uncertain whether cross-fostering contributed directly to later likelihood of tail-biting occurrence or whether this association was related to a common causal factor (for example, herd size or litter size). Hakansson and Houe (115) reported cross-fostering was associated with a higher probability of tail damage post-weaning. In contrast, Calderón Diaz et al. (275) found no influence of early or late cross-fostering on the occurrence of tail lesions. However, pigs from fostered litters were more at risk for death and euthanasia, with severe tail lesions being one of the reasons for euthanasia. Furthermore, cross-fostering placed pigs at greater risk of subsequent pericarditis and heart condemnations (275). A high level of cross-fostering is also a risk factor for PMWS (258).

In pens of weaners with high stocking density, prevalence of tail lesions was lower if total piglet losses were <18 % (12%, n = 45) and higher, if suckling piglet losses were ≥18% (26%, n = 116) (84). This might suggest that better health during lactation reduces later tail-biting risk.

In general, piglets weaned later are more developed and perform less manipulatory behavior after weaning (276–279), which might reduce DB. However, available data suggest that age at weaning (one to 6 weeks) had no clear influence on tail and ear biting in growing pigs even though manipulatory behavior was at least transiently increased in the piglets weaned at the earlier ages (one to 2 weeks compared to 4 weeks, but also 4 weeks compared to 6 weeks) [reviewed by (80)].

Early weaning stress (16 and 18 days of age vs. 20 days of age) resulted in immediate and long-term deleterious effects on intestinal defense mechanisms (280). Early weaning age [e.g., <26.5 days) increases the risk of developing PMWS (172, 281). Piglet age at weaning was also a risk factor for respiratory disease (121, 282). Moreover, the risk of lameness for pigs weaned at a younger age was higher (275).

The weaning period is associated with a lot of stress for piglets including removal from the sow, dietary changes, regrouping of pigs from different barns or even farms as well as adapting to a new environment. Indirect evidence suggests that increased weaning stress is associated with more DB. Treatments designed to facilitate the weaning transition, such as perinatal flavor conditioning (283, 284) or playful foraging during the lactation period (285) not only improved growth and feed intake post-weaning, but also reduced DB (248, 284) or number of pigs with ear or tail damage (285).

At weaning, histological changes in the small intestine have a negative effect on the immune system and lead to intestinal dysfunction which favors post-weaning diarrhea (286) and is aggravated by reduced feed intake. Hence, post-weaning alterations in gut functioning can increase disease susceptibility and mortality rates [reviewed by (287)].

Animal, feeding and environmental risk factors have a direct effect on the animal. In contrast, some of the risk factors associated with housing and management influence the animal, and therefore DB and disease, indirectly. For example, effects of farm size and biosecurity are probably mediated by related issues such as staffing level, hygiene practices and farm layout (288). Large farms are also more likely to have large group sizes. These in turn are often associated with more regrouping of pigs, although poor matching of pen sizes or uneven growth can also be reasons for repeated regrouping (128, 289). In any case, the direct detrimental effects on tail-biting and health of frequent regrouping, large group size and high stocking densities are consistent (Table 2). The same holds for the protective effects of solid floors and provision of manipulable materials. The benefits of bedding for tail biting and disease likely relate to better thermal and physical comfort and the fact that straw also acts a source of roughage. Indeed, the fact that pigs on slatted flooring do not have bedding and therefore a source of roughage, is a major reason why slatted floors are a risk factor for tail-biting. Straw bedding has some detrimental impacts on pig health but these are likely mediated by management factors such as poor hygiene and use of poor quality straw. Clearly, any housing and management practices that minimize weaning stress can simultaneously improve pig health and welfare and thereby the risk of disease and tail biting. Higher age at weaning is protective for pig health but while there are indications it is also protective for DB, the evidence is limited. Finally, aspects of lactation management such as cross-fostering are increasingly associated with later tail-biting (14) and poor health. However, it is uncertain whether cross-fostering contributes directly to later likelihood of problems, or whether the association is related to a common causal factor (for example, herd size or litter size).