Behavioural and physiological indicators of shelter ...

Physiology & Behavior 133 (2014) 223–229

Contents lists available at ScienceDirect

Physiology & Behavior journal homepage: www.elsevier.com/locate/phb

Behavioural and physiological indicators of shelter dogs' welfare: Reflections on the no-kill policy on free-ranging dogs in Italy revisited on the basis of 15 years of implementation S. Cafazzo a,b, L. Maragliano c, R. Bonanni a, F. Scholl d, M. Guarducci d, R. Scarcella e, M. Di Paolo d, D. Pontier f, O. Lai d, F. Carlevaro e, E. Bucci d, N. Cerini e, L. Carlevaro e, L. Alfieri d, C. Fantini c, E. Natoli c,⁎ a

Department of Neuroscience, University of Parma, Italy Wolf Science Center, Ernstbrunn, Vienna, Austria c Azienda USL Roma D, Rome, Italy d Istituto Zooprofilattico Sperimentale delle regioni Lazio e Toscana, Rome, Italy e Azienda USL Roma H, Rome, Italy f Laboratoire de Biometrie et Biologie Evolutive, Université de Lyon 1, Villeurbanne, France b

H I G H L I G H T S • • • • •

We assessed the welfare of dogs in 8 shelters in the Lazio Region (Italy). We utilised an innovative approach: oxidative stress and behavioural indicators. The variable that made the difference was to go out of the cage for a walk regularly. Dogs that enjoyed it had higher total antioxidant capacity and less anxious behaviour. Furthermore, they were more sociable to other dogs.

a r t i c l e

i n f o

Article history: Received 26 March 2014 Received in revised form 19 May 2014 Accepted 30 May 2014 Available online 4 June 2014 Keywords: Domestic dogs No-kill policy Oxidative stress·shelter Welfare

a b s t r a c t The Italian National Law 281 of 1991 forbids the euthanatization of free-ranging dogs, unless they have an incurable illness or are proved to be dangerous. Without neglecting the undeniable benefits of the “no-kill” policy, nevertheless it has brought about a chronic overpopulation in shelters and, as a result, higher costs of management and welfare problems since some dogs remain in the shelter for life. In 2004–2008, the Istituto Zooprofilattico Sperimentale of the Lazio and Tuscany regions carried out a survey in the Lazio Region to verify the effects of the Italian National Law 281/91 on free-ranging dog management following 15 years from its implementation. One of the aims of the study was an assessment of the welfare of dogs in a shelter sample (8 shelters out of 47 censused in the Lazio Region). 97 mixed-breed dogs were selected, their behaviour was studied and a blood sample was taken for each dog in order to determine the individual blood concentration of cortisol and the amount of oxidative damage (level of dRoms), as well as the amount of antioxidants to cope with it. Moreover, the total leukocyte count (leukogram) was accomplished. We ran general backward stepwise regression models using “level of antioxidant”, “level of dRoms” and “level of serum cortisol” as dependent variables respectively. The results showed that the most important variable that improved the level of welfare of dogs consisted in having the opportunity to regularly go out of the cage for a walk, whereas other variables like gender, size of the cage (small, medium, large), being alone in the cage, and being neutered/entire, had no significant effect on the physiological indicators of welfare. Dogs that enjoyed the regular walk had a higher total antioxidant capacity, and performed a lower frequency of displacing activities and stereotyped behaviour. Moreover, oxidative stress parameters seem to be indicators well matched with behavioural indicators of stress. Thus, for the first time, markers of oxidative status are utilised for the welfare evaluation in the domestic dog. Furthermore, the results of this paper give some suggestion about how small steps can help to improve shelters and, furthermore, this paper intends to solicit the debate on the no-kill policy. © 2014 Elsevier Inc. All rights reserved.

⁎ Corresponding author at: Azienda USL Rome D, Veterinary Hospital, via della Magliana 856, 00148 Rome, Italy. Tel.: +39 06 56487503. E-mail address: [email protected] (E. Natoli).

http://dx.doi.org/10.1016/j.physbeh.2014.05.046 0031-9384/© 2014 Elsevier Inc. All rights reserved.

224

S. Cafazzo et al. / Physiology & Behavior 133 (2014) 223–229

1. Introduction Despite a copious amount of research on animal welfare, surprisingly little is known about welfare indicators in shelter dogs. This is definitely not due to a lack of data (although, of course, new and scientific investigations of a high standard are always needed), but rather to the incredibly high number of variables that characterise the shelter environment and which make the identification of individual, as well as general, welfare indicators difficult (see for example [12]). For example, physiological parameters such as plasma, salivary and/or urinary cortisol concentrations are non-specific and may not reflect an emotional state, as reviewed by Taylor and Mills [33]. Furthermore, if we match such physiological parameters with dog behaviour, in order to look for correlates of welfare, the results are puzzling, as it has often been reported in the literature (reviewed in [30,33]). It would seem that behaviour does not reflect physiological stress in a consistent way (reviewed in [15]). For example, Hiby et al. [21] found only one relation between urinary cortisol levels and a behavioural pattern, out of 18 behavioural categories, in shelter dogs. Hennessy et al. [17] found no significant associations between plasma cortisol concentrations of shelter dogs and six behavioural factors derived by principal component analysis. Beerda et al. [4] found no significant correlation in kenneled dogs between urinary cortisol levels and six behavioural patterns they found to be characteristic of chronic housing stress. One of the reasons for the reported puzzling results might be that cortisol, especially plasma cortisol, is too dependent on the temporal context, which is why it is utilised as an acute stress indicator [15]. Some additional information can be yielded by the patterns of the individual leukogram, matched with the level of blood cortisol, because the leukogram provides valuable information regarding pathologic and physiologic responses to various stimuli [22]. It is well known that chronic stress may cause pervasive tissue damage and metabolic dysfunction, potentially leading to death. Metabolism produces pro-oxidant compounds capable of damaging biomacromolecules. Organisms cope with pro-oxidants by using endogenous and exogenous antioxidants that convert pro-oxidants into less reactive and damaging molecules [35]. Thus, a negative action of pro-oxidants is derived by an imbalance between reactive oxygen metabolite (ROM) production and neutralizing capacity of antioxidant mechanisms [32]. When this imbalance occurs, oxidative stress (OS) arises [14]. Several factors may affect the oxidative status of an individual. Among these, stress hormones that, if produced in excess, can increase OS [10,24,28,29]. Thus, OS could be utilised as a physiological indicator, as it has been proposed for swine [7]. Physiological welfare indicators, possibly matched with behavioural indicators, are indispensable tools in order to evaluate dog level of welfare and, moreover, they are crucial towards establishing the standards of management of dogs in shelters, especially in countries where the no-kill policy has been implemented by law. Among European countries, it is only in Italy, Austria and some Regions of Spain, where the law forbids the euthanasia of free-ranging dogs, unless they suffer from incurable illness or are proved to be dangerous. The National Italian Law 281 of 1991 makes the registration of owned dogs compulsory and supports spaying/neutering programmes, essential for prevention but which are not thoroughly enforced. Therefore, in spite of an average of 1000 adoptions per 1500 new intakes per year (regarding only the public dog shelter of Rome), the no-kill policy has brought about a chronic overpopulation in shelters and, as a result, has increased the costs of management and welfare problems since some dogs remain in the shelter for life. In 2004–2008, the Istituto Zooprofilattico Sperimentale of the Lazio and Tuscany regions carried out a survey in the Lazio Region to verify the effects of the Italian National Law 281/91 on free-ranging dog management, fifteen years after its implementation. One of the aims of the study was an assessment of the welfare of dogs in a shelter sample (8 shelters out of 47 censused in the Lazio Region). One part of

the survey was therefore designed to answer the question: ‘What is the quality of life of sheltered dogs living in the Lazio Region?’ This question necessarily posed, and poses, another question: ‘What do we know about the welfare indicators of sheltered dogs?’ (see for example [2–6, 16–20,30]). The aim of this study was then to assess the level of welfare of dogs housed in different kinds of shelters in the Lazio Region, utilising some behavioural and physiological measures, and to yield further information on their validity. Therefore, we tried to verify if oxidative stress can be utilised as welfare indicator of the domestic dog; in fact, OS has already been indicated as a cause of the progression of a chronic stressful state: prolonged high secretion of glucocorticoids in general, and of cortisol in particular, has been hypothesized to result in elevated oxidative stress [10]. Moreover, we aimed at comparing the physiological level of dogs' welfare/stress in shelters adopting different management methods, in order to detect variables that were likely to affect it significantly, in order to revisit the no-kill policy in Italy after 15 years of implementation.

2. Materials and methods 2.1. Animals and housing The subjects of this study were 97 mixed-breed dogs (52 males and 45 females) about half of which (40 animals) were neutered (19 males and 21 females). Dogs were selected among 3822 dogs that were housed in the shelters described below, on the basis of length of permanence in the shelter (2–3 years), age (2–7 years) and health conditions (unhealthy animals were excluded). The 97 dogs chosen included stray dogs (although socialised to humans), dogs relinquished to the shelters by the owners, and dogs sequestered by shelter staff because they were being abused; in other words, dogs commonly available at shelters for adoption. 77 dogs out of 97 were tested for a total of 385 h of behavioural observation (5 h/dog). The number of dogs observed in each shelter was chosen according to a random stratified sample (two levels: shelters and dogs) (Table 1). 7 out of 8 shelters were traditional shelters, i.e. dogs were housed in cages, alone or with some conspecifics. The 8th shelter was a dog park, where dogs live in big fenced areas. Given that in the dog park it was not possible to observe the dogs individually, 20 dogs were filmed. The films were then analysed and behavioural data collected using the same ethological methods mentioned above. It was not possible to include this data in the model given the different observational approach, but both physiological and behavioural data was useful in the comparison between shelters. Shelters were characterised by: i) the number of dogs present (smallb350 dogs; medium 350–700 dogs and large N700 dogs); ii) the presence/absence of volunteer dog care-takers. Where the volunteer dog care-takers were not present, dogs rarely went out of their cage for a walk. In some shelters there was a sort of compromise, i.e. the dogs were, in turn and periodically, placed in a larger cage to run freely, but alone. Dogs were classified as living in small (2–5 m2), medium (10–12 m2) or large (30 m2) cages; alone or with some conspecifics (range 1–8). Table 1 Shelters chosen and number of dogs observed in each shelter. Local health unit

Shelter

City and town

Province

No. of dogs observed

ASL RM D ASL RM E ASL RMF ASL RM G ASL RM H ASL FR ASL RI ASL LT

No. 1 No. 2 No. 3 No. 4 No. 5 No. 6 No. 7 No. 8

Roma Roma Bracciano Palestrina Pomezia Ceccano Rieti Fondi

RM RM RM RM RM FR RI LT

20 9 14 15 8 8 5 Filmed 20


2.2. Behavioural observations Each of the 77 dogs was observed in its cage from a distance that ranged from 2 to 5 m. All dogs were rapidly accustomed to the presence of the observer since the latter did not try to interact with them, neither during the observation session, nor between observation sessions: after few minutes they just went on doing their ordinary activities. The ethogram consisted of more than 100 behavioural patterns (described previously in [13]) (Table 2). Data was collected using the focal animal sampling method [1]. The observations were made with a check sheet recording the selected behavioural patterns of one dog (the focal animal) with ‘All occurrences’ and ‘1/0’ methods (60 s interval) [1] (see Table 2).

225

main feature, lymphopenia. Nevertheless other additional alterations are detectable such as monocytosis and an increase of mature neutrophils [22,23]. High level of dRoms indicates an extensive oxidative damage to biological molecules whereas the antioxidants are responsible for the delay or the inhibition of oxidative processes which occur under the influence of molecular oxygen or reactive oxygen species associated with the attack of free radicals. Reactive oxygen metabolites (dRoms) were determined by a colorimetric method (Diacron Adsorbent Test, Grosseto, Italy). The results of the dRoms test are expressed in arbitrary units called “Carratelli units” (CARR U), that is 1 CARR U = 0.08 mg of H2O2/100 ml. The Total Antioxidant Status test (Randox Lab. UK) is performed by a colorimetric method and the result is expressed in mmol/l.

2.3. Blood sampling and determination of physiological parameters A blood sample was taken from each dog in order to determine the individual blood concentration of cortisol (ng/ml), corrected for time of blood collection, and the oxidative status by the amount of oxidative damage (dRoms) as well as the total antioxidant capacity to cope with it [9]. All blood samples were kept refrigerated during the time necessary to take them to the laboratory. Serum was then separated by centrifugation and then frozen (−20 °C) until assayed. Concentrations of cortisol were determined in duplicate by radioimmunoassay using specific antibodies against cortisol, all with the same commercially available radioimmunoassay kit (Diasorin, Minnesota, USA). In 57 out of 77 dogs blood in an EDTA tube was collected and stored at 4 °C for the analysis of a total leukocyte count (leukogram) and the numbers of neutrophils, lymphocytes, eosinophils, basophils, and monocytes within 24 h after sampling (automated counter Cell Dyn 3700 — Abbott provided with veterinary software). The machine was calibrated daily with known controls (high and low) to ensure that all values are within a specified acceptable range. The total leukocyte count in healthy dogs varies from 5000 to 14,100 cells/μl blood (Duncan et al. 1994). The neutrophil is the most commonly encountered leukocyte in the blood of healthy dogs. Lymphocytes occur often but are less numerous than neutrophils. Monocytes and eosinophils are seen less frequently, whereas basophils are rare. Alterations in the total leukocyte count primarily are caused by changes in the numbers of lymphocytes and neutrophils. The assessment of the leukogram and its patterns is an excellent method to detect not only inflammation, severity of a disease or the response to a treatment, but also the occurrence of physical stresses or painful events. As a matter of fact, as a partial effect of increased cortisol level, the total number of circulating neutrophils becomes elevated while lymphocytes decrease (lymphopenia). This latter pattern is due to an excess of corticosteroid release causing, as the

2.4. Data analysis When possible, behavioural patterns have been gathered into categories (Table 2). Since there was a large number of variables and many of them were correlated, we applied a principal components analysis (PCA) in order to replace them with new uncorrelated component variables. The latter, called principal components or factors, were linear combinations of the original variables and explained most of the variation in the data. However, note that we have included in this analysis only those behavioural patterns that were displayed by at least two-thirds of the dogs under study (see Table 2). Moreover, despite the amount of literature available on displacing activities in domestic dogs (sometimes called ‘calming signals’), for this species there is not a uniform identification of the behavioural patterns that actually belong to this category. In contrast, a better uniformity of description of displacement activities occurs in other mammalian species (see for example [26]). Thus, including into the PCA displacing activities that express moderate levels of anxiety has also the function of clarifying this aspect of dog behaviour, given that it allows the assessment of the actual correlation between those behavioural patterns supposed to accomplish the same function (in this case to lower moderate level of anxiety). To explain individual variation in plasma oxidative balance, we ran general backward stepwise regression models (STATISTICA Release 8, StatSoft Inc., Tulsa, OK, U.S.A.) using “level of antioxidant”, “level of dRoms” and “level of serum cortisol” as dependent variables respectively. The predictor variables were: alone/with other dogs, type of walking out of the cage (i.e. walk out of the cage; placed in a larger cage alone; never out of the cage), neutered/entire, gender, age (scored as categorical variable), size of the cage (small, medium, large) and the first six

Table 2 Categories of behavioural patterns utilised in the study (see De Palma et al. 2005). Behavioural categories

Behavioural patterns

Displacement activities Stereotyped or repetitive behaviour Agonistic behaviour

Body shaking°*, scratching°*, muzzle licking°*, yawning°*, turning on themselves°*; autogrooming+*. Repetitive pacing in circles°*, licking or biting bars and object repeatedly+, catching “invisible flies”°*, coprophagy°*, self mutilation+*. Aggressive behaviour: growling°, transverse glance°, raising fur°, curling lip°, showing teeth°, biting bars°, dashing at bars°; dominance behaviour: staring°, tail still°, stiff body°, raised tail+, wagging with the tail held high+; pricked-up ears°; putting the paw or a muzzle on a conspecific's back°; submissive behaviour: avoiding eye contact°*, lowering head°*, flattening ears°*, cringing°*, holding the tail down or tightly between the hindlegs and against the belly°*, laying down on the back exposing the ventral side of the chest and sometimes the abdomen°*. Tail wagging+*, giving the foreleg°, leaning on bars°, licking the mouth of other dogs°*, passive contact+*, allogrooming+*. Walking°*, trotting°*, galloping°*. Standing°*, sitting°*, laying°*. Dozing (appears near sleep)°*. Raising ears°*, looking outside the cage carefully°*, looking at the observer carefully°*, looking at another dog carefully°*, raising one fore leg°*, raising both fore legs on the wall or on the bars°*. Looking outside the cage°*, looking at the observer°*, at another dog°* Smelling their environment°*, smelling the observer°*, another dog°*. Inviting to play°, answering inviting to play°, showing object°. Barking+*, whining°, grumbling°, mumbling°, howling°, snorting°.

Affiliative behaviour Locomotor activity Change of posture Resting Total attention Total looking Olfactory investigation Playing Vocal communicability

°Behavioural patterns scored by all occurrences method; +behavioural patterns scored by one-zero method; *behavioural patterns involved in the PCA.


factors of the PCA. Model residuals were tested for normality using the Kolmogorov–Smirnov test. We also applied a one-way ANOVA with ‘level of antioxidant’, ‘level of dRoms’ and ‘level of serum cortisol’ as a dependent variables, respectively, and ‘presence/non-presence of volunteer dog care-takers’ as a factor. 3. Results The first six factors of the PCA taken together explained about 70% of the total variance in the data. The first factor of the PCA had a high negative correlation with measures of attention, looking out of the cage, movement (walking, trotting, galloping) barking, turning on themselves, sniffing the environment and submissive behaviour towards humans. The second factor was negatively correlated especially to variables supposed to function as displacement activities such as autogrooming, scratching and body shaking; moreover it was positively related to dozing and to total affiliative behaviour displayed towards other dogs. The third factor was positively correlated to yawning and to change of position (standing, sitting and laying); moreover, this factor was also negatively correlated to stereotypical behavioural patterns such as repetitive pacing in circles, catching flies, coprophagy, the act of scratching the platform and/or the ground. The fourth factor was positively related to protrusion of the tongue to lick the lips (possibly a displacement activity) and negatively related to sniffing the ground. The fifth factor had a negative correlation with spinning around. Finally the sixth factor was positively correlated to wagging the tail towards humans (Table 3). The general linear model developed for the ‘level of antioxidant’ was highly significant and showed a very good fit (F6,68 = 17.55, P = 0.0000001; R = 0.78). The most significant predictor of the dependent variable was “walk out of the cage” which had a positive effect (coefficient ± SE = 425.84 ± 60.47, t = 7.04, P = 0.0000001). In other words, dogs that had the opportunity to go out of the cage for a walk regularly had a higher level of antioxidant capacity than dogs that never went out, as well as than dogs that were placed in a larger cage, but alone (Fig. 1). Conversely, dogs that never went out of the cage displayed a level of antioxidant capacity that was not significantly different from that of dogs placed in a larger cage alone (t = − 0.64, P = 0.52; see Fig. 1). Other significant predictors were ‘Factor 2’ and ‘Factor 3’ of the PCA, both of which had a positive effect on the antioxidant capacity (Factor 2: coefficient ± SE = 65.41 ± 29.82, t = 2.19, P = 0.03; Factor 3: coefficient ± SE = 80.62 ± 29.94, t = 2.69, P b 0.009). So, the model suggests that dogs that rarely showed autogrooming, scratching, body

Mean values of antioxydants (µmoli/l)

226

3000 2500 2000 1500 00 100 0 00 50 0

0 Go out

Do on't go o ou ut

er I a larrge In un ge to t ru c cag e one alo

Fig. 1. Level of antioxidant capacity in dogs that went regularly out of the cage for a walk, never went out, and that were placed in a larger cage, but alone (mean values ± SE).

shaking behaviour, and that frequently dozed and displayed affiliative behaviour to other dogs had a higher level of antioxidant capacity. Moreover, dogs that rarely showed stereotypical behavioural patterns and frequently showed yawning, frequent changing of posture and dozing also had a higher level of antioxidant capacity. Altogether, the model suggests that the frequent display of behaviours such as autogrooming, scratching, body shaking and stereotypic patterns may indicate a lower level of antioxidant capacity in shelter dogs. We also found that antioxidants were affected by age, meaning that dogs belonging to the oldest age class had higher antioxidant capacity than dogs of intermediate age (coefficient ± SE = 268.02 ± 82.92, t = 3.23, P b 0.002), and younger dogs had lower antioxidant capacity than dogs of intermediate age (coefficient ± SE = − 318.46 ± 72.32, t = −4.40, P b 0.00004). Conversely, it was not possible to build a model neither for the ‘level of dRoms’ nor for the “level of cortisol” because in these cases the backward stepwise procedure did not detect any significant predictor variable. However, male dogs tended to have higher levels of dRoms than females (t = 1.96, P = 0.054), and dogs of the oldest age class tended to have lower dRoms levels than dogs of the intermediate age class (t = −1.71, P = 0.09). The opportunity to go out of the cage regularly for a walk is due to the presence of volunteer dog care-takers. Consequently, their presence influenced the ‘level of antioxidant’ that was higher when they were present (ANOVA: F(1,95) = 61.25, P = 0.000001) (Fig. 2a) but it did not affect the ‘level of dRoms’ (ANOVA: F(1,95) = 1.45, NS) (Fig. 2b) and blood cortisol level, although there was a tendency to be lower when they were present (ANOVA: F(1,91) = 3.45, P = 0.07) (Fig. 2c).

Table 3 Correlations between the first six factors of the PCA and the behavioural variables under investigation. Bold-faced data marks correlations≥0.4.

Autogrooming Total stereotyped behaviour Turning on themselves Yawning Self muzzle-licking Scratching Body shaking Barking Total submissive behaviour to humans Tail wagging to humans Total affiliative behaviour to dogs Total locomotory activity Total changing posture Dozing Total looking Total attention Total olfactory investigation Autogrooming

Factor 1

Factor 2

Factor 3

Factor 4

Factor 5

Factor 6

0.308 −0.243 −0.487 −0.230 −0.352 0.352 0.202 −0.644 −0.588 −0.364 0.175 −0.798 −0.319 0.127 −0.774 −0.867 −0.488 0.308

−0.605 −0.298 0.151 0.156 −0.141 −0.662 −0.639 −0.198 −0.123 0.149 0.490 −0.196 0.021 0.563 −0.006 0.010 −0.178 −0.605

0.143 −0.474 −0.119 0.636 0.222 0.285 0.307 −0.425 −0.226 0.076 −0.033 0.230 0.530 0.494 0.094 −0.031 0.395 0.143

0.213 −0.135 −0.106 −0.231 0.611 −0.165 −0.155 0.022 −0.101 0.485 −0.240 0.060 0.415 −0.083 −0.205 −0.023 −0.589 0.213

−0.163 0.262 −0.673 0.076 −0.223 −0.123 0.062 0.341 0.425 −0.065 0.014 0.053 0.436 0.203 −0.040 −0.346 −0.049 −0.163

0.072 0.461 −0.027 0.204 −0.359 0.053 −0.057 −0.009 −0.367 0.638 0.150 0.227 −0.014 −0.082 −0.329 0.029 0.138 0.072

a

200 00 2 0 150 00 0 100 00 0 00 0 50 0

a 25 500 0.0 20 000 0.0 15 500 0.0 10 000 0.0

Y Yess N No nce e o of vollun are Pressen nteer do og ca e ta ake erss

5 500 0.0 0 0.0

b

40

Dog gs she elte ers s

35 30

60 0.00

25 20 15 10 5 0 Y Yess N No Pressen nce nteer do og ca e ta ake erss e o of vollun are

3..00 0

Mean level of dRoms (U.Carr)

M Mean llevell of f dR dRoms (U (U.Carr) C )

45

b

50 0.00 40 0.00 30 0.00 20 0.00 10 0.00 0 0.00

c

2..50 0 2..00 0

Do og sh helte erss

1..50 0 1..00 0 0..50 0 0..00 0

Y Yess N No Pressen nce e o of vollun are nteer do og ca e ta ake erss

Fig. 2. Level of antioxidant capacity (a), dRoms (b) and serum cortisol (c) in dogs living in shelters with and without volunteer dog care-takers (mean values ± SE).

These results were reflected in the comparison among shelters: where dogs never went out of the cage because of management choices and absence of volunteer dog care-takers, ‘level of antioxidant’ was lower (ANOVA: F(7,89) = 47.64, P = 0.00001) (Fig. 3a) although ‘level of dRoms’ did not differ significantly (ANOVA: F(7,89) = 2.08, NS) (Fig. 3b) and level of blood cortisol was moderately higher (ANOVA: F(7,85) = 1.84, P = 0.09) (Fig. 3c).

Mean level of serum corsol (ug/dl) (corrected for me of ll i collection

Mean level M l l of f serum cortisol ti l (corrected for time (ug/dl) g of collection

227

30 000 0.0

250 00 2 0

Mean values of antioxidants (µmol/l)

Mean values of antioxidants (µmoli/l)


5.0 00 4.5 50 4.0 00 3.5 50 3.0 00 2.5 50 2.0 00 1.5 50 1.0 00 0.5 50 0.0 00

c

Do og sh heltterrs Fig. 3. Comparison among shelters with and without volunteer dog care-takers for level of antioxidant capacity (a), dRoms (b) and serum cortisol (c) (mean values ± SE). Volunteer dog care-takers were present in the first three shelters from left, labelled by an asterisk.

3.1. Blood parameters and physiological indicators Concerning blood parameters, although neutrophils were correlated positively to monocytes (r = 0.60, N = 57, P = 0.0001) and basophils (r = 0.64, N = 57, P = 0.0001), there was neither positive correlation with blood cortisol nor negative correlation with lymphocytes. Thus, their values were not attributable to a canine stress leukogram pattern. 4. Discussion The results of this study clearly suggest that the most important variable that improved the level of welfare of dogs was having the opportunity to regularly go out of the cage for a walk. In fact, dogs that had regular walks had a higher total antioxidant capacity than dogs that did not. Moreover, our results also suggest that dogs that performed a lower frequency of displacing activities

(particularly autogrooming, scratching and body shaking) and stereotyped behaviour (particularly repetitive pacing in circles, catching flies, coprophagy, the act of scratching the platform and/or the ground) had a higher total antioxidant capacity. Notably, dogs that seemed more relaxed because they were more sociable to other dogs and often rested, changing position (standing, sitting and lying) but dozing in all three positions, had a higher total antioxidant capacity and thus a better level of welfare. In the context of this study, it is difficult to distinguish if the higher level of welfare was due to the daily walk or to the daily interaction with human beings, or to both things. Nevertheless, it is to underline that dogs that were periodically put in larger cages to roam freely, but alone, had levels of total antioxidant capacity and frequency of some displacing activities (autogrooming, scratching and body shaking) comparable to those dogs that never went out for a walk. These results give

228


indication that, confirming the results of Coppola et al. [8], Hennessy et al. [18–20], McGreevy et al. [27], and Tuber et al. [34], reviewed in Hennessy [15], the interaction with human beings is undoubtedly important to improve dog welfare. In the comparison among shelters this trend is even clearer (Fig. 3): the presence of volunteer dog caretakers in the shelter makes the difference. As has often been reported in the literature (reviewed in [30,33] and, recently, in [15]), the results of this study confirm that the relationships between hematic cortisol and stress-induced behaviour (both displacing activities and stereotyped behaviour) remain unresolved. Although glucocorticoid concentrations in general, and cortisol in particular for dogs, are the most commonly used measure of welfare, Hennessy [15] underlined that the results concerning it have to be considered cautiously. This is because, first of all, cortisol is only one element of a complex system and, secondly, it is particularly useful for evaluating acute stress. Consequently, the lack of correspondence between levels of cortisol and stress-induced behaviour that has been found in this study suggests that dogs were experiencing a state of chronic stress, because they had been housed in the shelter for a minimum of two years. It is not to be forgotten that chronic stress may itself deregulate the entire system [25]. This can be also one of the reason why, in this study, there was no any relationship found between cortisol and blood elements like neutrophils, monocytes, basophils and lymphocytes, attributable to a canine stress leukogram pattern. On the contrary, the measures of oxidative stress have been shown to be less dependent on temporarily acute stress although, in longer terms, stress response mediated by glucocorticoids may increase the oxidative stress (OS) in both birds and rats (e.g. [10,24,31]). In fact, organisms have evolved more stable multiple lines of defense to prevent oxidative damage due to acute stress, ranging from antioxidant enzymes to low molecular weight antioxidants, to specific cellular components that repair oxidatively damaged molecules [9]. Despite the fact that information on antioxidant capacity by itself is not sufficient to make inferences about oxidative stress [11], high levels of antioxidants undoubtedly help in a stress condition. Our results suggest that this high level corresponds to low frequency of anxious and disturbed behaviour and to higher frequency of affiliative behaviour, and that all these parameters are linked to the possibility of having a regular contact with human beings and of going regularly out of the cage for a walk, or to both together. So, since the latter is linked to the presence of volunteer dog care-taker in the shelters, the results also suggest that the presence of the volunteer dog care-takers in the shelter makes a clear difference: when they are absent, dogs showed signs of lower welfare. Consequently, despite the fact that the presence of Associations of Volunteer Dog care-takers sometimes brings about management conflicts due to their different cultural backgrounds in respect to the cultural background of health teams (veterinary doctors, biologists, ethologists, nurses, etc.) their presence improves the quality of life of shelter dogs. 5. Conclusions Our results suggest that the utilisation of oxidative stress parameters as welfare indicators of the domestic dog might be successful. In this study, differences in physiological status of OS were also coupled with behavioural responses, such as increase in anxiety (displacing activity and stereotyped behaviour). So far, there is scant information on the relationship between behaviour and oxidative status. Our results suggest also other, more general considerations: fifteen years of implementation of the no-kill policy on free-ranging dogs in Italy cannot be considered particularly successful because: • in terms of free-ranging dog population management, the law has not resolved the problem because shelters are not able to take in exhaustively the unwanted free-ranging dog population • in terms of dog welfare, the law does not enforce creation of conditions in the shelters to look after a dog for life properly.

For example, in 2005 only in the Lazio Region there were over 11,000 dogs censused in private and public shelters, and many of them stayed there for their entire lives. Given that, in Italy, we have decided for dogs that life-imprisonment is better than painless euthanasia, it is our ethical duty to guarantee them an adequate level of welfare. It is evident from the scientific literature that this is not the case. The results of this paper give some suggestion about how small steps can help to create, if not an optimal shelter (it does not exist), at least an improved shelter. In our opinion, the no-kill policy deserves a wider debate about costs of management and the level of welfare of sheltered dogs, based on data coming from countries where it has been enforced by law.

Acknowledgements We wish to thank the shelter staff at all shelters involved for their continual cooperation in conducting this research; in particular, our thanks go to the vets for their help and assistance to collect the blood samples during the study. Our thanks go also to Mario Di Traglia for the assistance during statistical analysis. The project was financially supported by The Italian Minister of Health, Ricerca Corrente IZSLT 09/ 05 RC. Writing was supported by FWF Lise Meitner Position (project number: M1400-B19) to the first author.

References [1] Altman SA. Observational study of behaviour: sampling methods. Behaviour 1974;49:227–67. [2] Beerda B, Schilder MBH, van Hooff JARAM, de Vries HW. Manifestations of chronic and acute stress in dogs. Appl Anim Behav Sci 1997;52:307–19. [3] Beerda B, Schilder MBH, van Hooff JARAM, de Vries HW, Mol JA. Behavioural, saliva cortisol and heart rate responses to different types of stimuli in dogs. Appl Anim Behav Sci 1998;58:365–81. [4] Beerda B, Schilder MBH, van Hooff JARAM, de Vries HW, Mol JA. Chronic stress in dogs subjected to social and spatial restriction. I. Behavioral responses. Phys Behav 1999;66:233–42. [5] Beerda B, Schilder MBH, Bernardina W, van Hooff JARAM, de Vries HW, Mol JA. Chronic stress in dogs subjected to social and spatial restriction. II. Hormonal and immunological responses. Phys Behav 1999;66:243–54. [6] Beerda B, Schilder MBH, van Hooff JARAM, de Vries HW, Mol JA. Behavioral and hormonal indicators of enduring environmental stress in dogs. Anim Welf 2000;9:49–62. [7] Brambilla G, Civitareale C, Ballerini A, Fiori M, Amadori M, Archetti LI, et al. Response to oxidative stress as a welfare parameter in swine. Redox Rep 2002;7:159–63. [8] Coppola CL, Grandin T, Enns RM. Human interaction and cortisol: can human contact reduce stress for shelter dogs? Phys Behav 2006;87:537–41. [9] Costantini D, Carere C, Caramaschi D, Koolhaas JM. Aggressive and non-aggressive personalities differ in oxidative status in selected lines of mice (Mus musculus). Biol Lett 2008;4:119–22. http://dx.doi.org/10.1098/rsbl.2007.0513. [10] Costantini D, Marasco V, Møller AP. A meta-analysis of glucocorticoids as modulators of oxidative stress in vertebrates. J Comp Physiol B 2011;181:447–56. http:// dx.doi.org/10.1007/s00360-011-0566-2. [11] Costantini D, Verhulst S. Does high antioxidant capacity indicate low oxidative stress? Funct Ecol Forum 2009:1–4. http://dx.doi.org/10.1111/j.1365-2435.2009.01546.x. [12] Dawkins MS. A user's guide to animal welfare science. Trends Ecol Evol 2006;21(2):77–82. [13] De Palma C, Viggiano E, Barillari E, Palme R, Dufour A, Fantini C, et al. Evaluating the temperament in shelter dogs. Behaviour 2005;142(9–10):1313–34. [14] Finkel T, Holbrook NJ. Oxidants, oxidative stress and the biology of ageing. Nature 2000;408:239–47. [15] Hennessy MB. Using hypothalamic–pituitary–adrenal measures for assessing and reducing the stress of dogs in shelters: A review. Appl Anim Behav Sci 2013;149:1–12. http://dx.doi.org/10.1016/j.applanim.2013.09.004. [16] Hennessy MB, Davis HN, Williams MT, Mellott C, Douglas CW. Plasma cortisol levels of dogs at a county animal shelter. Phys Behav 1997;62:485–90. [17] Hennessy MB, Voith VL, Mazzei SJ, Buttram J, Miller DD, Linden F. Behaviour and cortisol levels of dogs in a public animal shelter, and an exploration of the ability of these measures to predict problem behaviour after adoption. Appl Anim Behav Sci 2001;73:217–33. [18] Hennessy MB, Voith VL, Hawke JL, Young TL, Centrone J, McDowell AL, et al. Effects of a program of human interaction and alterations in a diet composition on activity of the hypothalamic–pituitary–adrenal axis in dogs housed in a public animal shelter. J Am Vet Med Assoc 2002;221(1):65–71. [19] Hennessy MB, Voith VL, Young TL, Hawke JL, Centrone J, McDowell AL. Exploring human interaction and diet effects on the behavior of dogs in a public animal shelter. J Appl Anim Welf Sci 2002;5(4):253–73.

S. Cafazzo et al. / Physiology & Behavior 133 (2014) 223–229 [20] Hennessy MB, Williams MT, Miller DD, Douglas CW, Voith VL. Influence of male and female petters on plasma cortisol and behaviour: can human interaction reduce the stress of dogs in a public animal shelter? Appl Anim Behav Sci 1998;61:63–77. [21] Hiby EF, Rooney NJ, Bradshaw JWS. Behavioural and physio-logical responses of dogs entering re-homing kennels. Physiol Behav 2006;89:385–91. [22] Jain NC. Schalm's Veterinary Hematology. 4th ed. Philadelphia: Lea & Febiger; 1986. [23] Latimer KS, Rakich PM. Clinical interpretation of leukocyte responses. Vet Clin North Am Small Anim Pract 1989;19:637–68. [24] Lin H, Decuypere E, Buyse J. Oxidative stress induced by corticosterone administration in broiler chickens (Gallus gallus domesticus) 1. Chronic exposure. Comp Biochem Physiol B 2004;139:737–44. [25] Korte SM. Corticosteroids in relation to fear, anxiety and psychopathology. Neurosci Biobehav Rev 2001;25:117–42. [26] Maestripieri D, Schino G, Aureli F, Troisi A. A modest proposal: displacement activities as an indicator of emotions in primates. Anim Behav 1992;44:967–79. [27] McGreevy PD, Righetti J, Thomson PC. The reinforcing value of physical contact and the effect on canine heart rate of grooming in different anatomical areas. Anthrozoos 2005;18:236–44. [28] Ohtsuka A, Kojima H, Ohtani T, Hayashi K. Vitamin E reduces glucocorticoid-induced oxidative stress in rat skeletal muscle. J Nutr Sci Vitaminol 1998;44:779–86.

229

[29] Rajashree S, Puvanakrishnan R. Dexamethasone induced alterations in enzymatic and nonenzymatic antioxidant status in heart and kidney of rats. Mol Cell Biochem 1998;181:77–85. [30] Rooney NJ, Gaines SA, Bradshaw JWS. Behavioural and glucocorticoid responses of dogs (Canis familiaris) to kennelling: investigating mitigation of stress by prior habituation. Phys Behav 2007;92:847–54. [31] Sahin E, Gumuslu S. Stress-dependent induction of protein oxidation, lipid peroxidation and anti-oxidants in peripheral tissues of rats: comparison of three stress models (immobilization, cold and immobilization-cold). Clin Exp Pharmacol Physiol 2007;34:425–31. http://dx.doi.org/10.1111/474 j.1440-1681.2007.04584.x. [32] Sies H. Oxidative stress II. Oxidants and Antioxidants. London: Academic Press; 1991. [33] Taylor KD, Mills DS. The development and assessment of temperament tests for adult companion dogs. J Vet Behav 2006;1:94–108. [34] Tuber DS, Hennessy MB, Sanders S, Miller JA. Behavioral and glucocorticoid responses of adult domestic dogs (Canis familiaris) to companionship and social separation. J Comp Psychol 1996;110:103–8. [35] Yu PB. Cellular defenses against damage from reactive oxygen species. Physiol Rev 1994;74:139–62.