Physiology & Behavior 87 (2006) 537 541

Human interaction and cortisol: Can human contact reduce stress

for shelter dogs?
Crista L. Coppola *, Temple Grandin, R. Mark Enns
Department of Animal Sciences, Colorado State University, Fort Collins, Colorado, 80523, USA

Received 3 March 2005; received in revised form 2 December 2005; accepted 6 December 2005

Abstract

Animal shelters are an extremely stressful environment for a dog, most specifically due to social isolation and novel surroundings. The stress
response of dogs housed in this environment may be alleviated through human interaction shortly after arrival. During their second day in a public
animal shelter, adult stray dogs were either engaged in a human contact session or not. The session involved taking the dog into an outdoor
enclosure, playing with the dog, grooming, petting and reviewing basic obedience commands. Each dog interacted with a human for
approximately 45 min. Salivary cortisol levels were examined from each dog on their 2nd, 3rd, 4th and 9th day of housing. Animals that engaged
in a human contact session had lower cortisol levels on day 3 than animals that did not. Breed type, sex and age did not have an effect on cortisol
levels on any day measured. A human interaction session can be beneficial to both animal welfare and adoption procedures. The current study not
only utilized the human contact session as a treatment to reduce stress but also as a resource for individual temperament/personality information
that could be later used to facilitate compatible adoptions. Human interaction may be an effective means of reducing the cortisol response of dogs
in the aversive shelter environment.
D 2005 Elsevier Inc. All rights reserved.

Keywords: Animal shelter; Animal welfare; Dog; Human interaction; Salivary cortisol; Stress response

1. Introduction shelter environment, intermediate levels on days 4 9, and to

Psychological stressors are the most potent natural stimulus
known to affect the pituitary adrenal cortex [1]. Dogs are
affected by many of these psychological stressors, most notably
novel, unpredictable or uncontrollable situations. Some specific
stressors that influence the hypothalamic pituitary adrenal
(HPA) axis include exposure to novel or threatening surround-
ings [2], separation from attachment objects [3], unpredictabil-
ity of external events [4], and lack of or loss of control over the
environment [5] with the most influential for dogs being novel
surroundings and separation from social attachments [3]. When
dogs are housed in an animal shelter they potentially
experience a severe form of social isolation and novel
surroundings [6]. In two studies [3,6], dogs were found to
have increased cortisol levels during the first three days in a
* Corresponding author. Present address: ASPCA Animal Behavior Center,
1717 S. Philo Road Suite 36 Urbana, IL 61802, USA. Tel.: +1 217 337 9785.
E-mail address: cristac@aspca.org (C.L. Coppola).
0031-9384/$ - see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.physbeh.2005.12.001
plateau at lower levels around day 9 10 (above household
pets). During the first three days in the shelter environment
dogs have cortisol levels that are almost three times that of
normal household dogs [7].
In dogs, stress response to aversive stimuli can be controlled
by the ability of the individual to control or predict the situation
[8]. Some interactions with humans might also be capable of
alleviating this stress response [3,6] and simple petting has
been found to eliminate stress from an electric shock [9].
Cortisol is recognized as the major indicator of response to
stress [10,11] and is used in animal welfare studies to detect poor
welfare [12 14]. Salivary cortisol is a better measure of adrenal
cortical function and is a better physiological indicator of stress
than plasma cortisol [10,15,16] because it is a direct reflection of
the biologically active portion of the total cortisol level [17]. The
measurement of salivary cortisol has been used as a noninvasive
method for measuring stress in dogs [14,18,19].
In dogs, there are no documented differences in basal levels
of cortisol between sexes [20,21,10]. Basal cortisol levels are
influenced by size and age but not reproductive status [20,21].
538 C.L. Coppola et al. / Physiology & Behavior 87 (2006) 537 541
Smaller breeds tend to have higher basal cortisol levels than
larger breeds and nursing pups have higher levels than older
dogs [20,21,6]. Stress induced cortisol levels do not follow the
same trends. In stress situations, females have higher cortisol
levels than males [22] and juvenile/pups have lower cortisol
levels than adults [7].
The purpose of this study was to examine the effect a human
contact session implemented shortly after arrival would have
on the stress response of adult dogs housed in a public animal
shelter. This study is one of the first to demonstrate that human
contact can reduce the stressful effects of shelter housing.
2. Materials and methods
2.1. Subjects and housing
This study protocol followed all USDA Animal Welfare
Regulations and was approved by Colorado State Universitys
Animal Care and Use Committee. The subjects were adult dogs
(as verified by absence of deciduous teeth) of various and mixed
breeds. The dogs were housed at a local county humane society,
a large rural shelter serving 3 municipalities and 15 townships.
All dogs arriving at the shelter were, by law, kept a minimum of
3 days before being made available for adoption. Subjects were
selected from among all those brought in as strays, were
apparently healthy, and did not appear aggressive (i.e. dogs that
growled, snarled or attempted to bite the handler were not
eligible). If there were small dogs and large dogs eligible to
participate on the same day, they were first separated by size
(small or large) and then randomly assigned to each study
group. If there were too many dogs eligible to participate on one
single day, due to time constraints, the maximum number of
dogs was randomly chosen from all individuals eligible.
Dogs were initially housed individually in kennels in one of
two stray rooms until they were available for adoption. Large
stray dogs were housed in a room (45 m2) with 15 runs
separated by cement block partition walls (2.18 m high),
concrete floors, concrete walls and an exposed rafter ceiling
(4 m high). Small stray dogs were housed in a room (24 m2)
with a bank of metal cages [7] of different sizes along one wall
and a double row of freestanding cages [10] in the middle of
the room. The floor and walls were concrete and the ceiling
was a suspended drop ceiling (2.44 m high).
Once a dog was eligible for adoption ( 3 days) and a space
was available in the corresponding adoption area the dog was
moved. Large adoptable dogs were housed in individual runs in
a freestanding room (81.75 m2) surrounded by an 8V cement
block perimeter wall. Dog runs [26] lined each of the four
perimeter walls and were separated by cement block partition
walls (2.10 m high). The floors were concrete and the ceiling
(6.10 m high) consisted of exposed rafters. Small adoptable
dogs were housed in individual cages in a room (26.50 m2)
with 28 metal cages of different sizes along three walls. The
floor and walls were concrete and the ceiling was a suspended
drop ceiling (2.44 m high).
All rooms were illuminated with a combination of overhead
fluorescent and natural lighting. The lights were manually
turned off at night (2200 0500 h). The number of dogs in the
small rooms varied slightly while the large rooms usually
remained at full capacity.
2.2. Experimental design
This experiment used a cross-sectional design to assess the
salivary cortisol levels of dogs housed in the shelter that had
engaged in a human contact session shortly after arrival (day
2). This experiment also used a longitudinal design to assess
salivary cortisol levels of dogs housed in the shelter on days 2,
3, 4 and 9, where day 1 was day of arrival.
There were two treatments: treatment 1, contact (n = 68),
contained animals that had engaged in a human contact session
on day 2 as described below and treatment 2, no-contact
(n = 62), included animals that had no such contact.
2.3. Human contact session
On each study day, potential subjects were first identified
and then randomly assigned to each treatment, contact (human
contact session) and no-contact. Animals in the contact group
engaged in a human contact session on day 2. Human
interaction with the dog took place in their kennel, outside in
a fenced enclosure (18V  12V) and inside in a small visitation
room (8V6VV  9V4VV). The contact session included playing with
the dog outside, walking on leash, grooming activities, playing
inside with toys, obedience commands, treat rewards, and
verbal and tactile contact. Each session lasted an average of
45 min (30 90 min) and was conducted by the same female
human evaluator. The sessions varied in length because of
individual animal variation in receptiveness, fear, socialization
and playfulness.
Stress response was measured using salivary cortisol levels.
Saliva samples were obtained from contact animals on day 2 at
least 2 h after the human contact session and from the no-
contact animals on day 2. If animals were still present in the
shelter on days 3, 4 and 9 subsequent saliva samples were
collected (Table 1). One hundred and five dogs had saliva
collected and analyzed on day 2, 92 on day 3, 83 on day 4 and
61 on day 9. A total of 55 dogs (25 Contact and 30 No-contact)
had saliva taken on all 4 sample days. Animals would not have
a saliva sample if the sample was physically impossible to
obtain (i.e. uncooperative, dog not available, not enough
sample collected).
2.4. Saliva collection and hormone determination
Saliva samples (1.0 ml) were collected using a cotton bud
with a string attached. The bud was sprinkled with citric acid
Table 1
Number of dogs (saliva samples) in each group on each sample day
Day 2 Day 3 Day 4 Day 9
Contact 54 48 38 28
No-contact 51 44 45 33
 C.L. Coppola et al. / Physiology & Behavior 87 (2006) 537 541
and then placed in the dogs mouth for up to 60 s. The wad was
removed and placed in a 10 cc syringe. The saliva was
removed from the wad by squeezing the contents of the syringe
into a 3 cc plastic micro-centrifuge tube. Dogs were sampled in
the morning hours (0900 1200) either inside their kennel or
just outside their kennel door. All samples were collected
within 5 min of the onset of disturbance (e.g. opening of
animals kennel). Based on previous studies, this short time
period would not affect cortisol levels in the saliva [23 25]. In
addition, as is noted in previous research, the initial attention
given to the animal during collection of saliva was most likely
pleasant rather than stressful [6].
Samples were stored at 20 -C until time of assay. Saliva
samples were thawed at room temperature (25 -C) and
duplicate aliquots were assayed for cortisol using an enzyme
immunoassay kit (Activei Cortisol EIA, Diagnostic Sys-
tems Laboratories, Webster, Texas). The assay was manufac-
tured for human saliva and was validated in our lab for canine
use (data not presented). The modifications to the assay
included extraction of the sample with dichloromethane and
reconstitution with a neutral (7.2 pH) buffer solution to a 4
concentration of the original sample. The modified assay
sensitivity was 0.011 Ag/dl. Intra- and inter-assay coefficients
of variation were 4.8% and 1.2%, respectively. Recovery was
93.3%.
2.5. Statistical analysis
The salivary cortisol data were analyzed using mixed model
procedures appropriate for repeated measures using Proc
Mixed in SAS (2002). Cortisol levels were log transformed
for analyses. In preliminary analyses the data were analyzed
with a statistical model that included the fixed effects of
treatment (contact or no-contact), sex (male or female), time
(day of sample), all two and three way interactions of the
former, breed type (work, herd, sport, nonsport and hound as
classified by the American Kennel Club), and age (estimated
by dentition in months) as a covariate. Animal within treatment
was included as a random effect. An unstructured correlation
0.6
0.5
Salivary Cortisol (ug/dl)
0.4
0.3
0.2
0.1
0
2 3*
*P=0.021
Fig. 1. LS Mean salivary cortisol levels of dogs in the Contact and No-contact treatments. Vertical lines represents 95% confidence intervals.
539
structure allowing for heterogeneous variances was used to
model the within-subject errors and the Kenwardrogers
approximation was used to estimate denominator degrees of
freedom. Preliminary analyses indicated that breed type, age
and the interaction of treatment, sex, and breed type were not
significant. These effects were removed from the final model.
Results presented are from this reduced model.
3. Results and discussion
Results shown are on the observed scale with significance
values based on analysis of transformed data. There was not a
significant treatment effect; however the interaction between
treatment and time ( F(1, 72) = 3.34, P = 0.024) indicated that
treatment effects varied over time. The greatest difference
between treatment levels occurred on day 3 (P = 0.021). As
seen in Fig. 1, there was an overall tendency for contact dogs to
have lower cortisol levels than no-contact dogs on day 3. There
were no significant differences between treatments on any
other day. Age, sex and breed type were not found to have a
significant effect on cortisol level.
Previous studies have indicated human interaction may
alleviate some stress for dogs housed in a shelter environment
[6]. In addition, many of the problems animals face in the
shelter can be alleviated through daily human contact and basic
conditioning to the stressful environment [26]. In the present
study, a human contact session of approximately 45 min
implemented on the 2nd day in the shelter was influential in
reducing stress response on the following day (day 3). Day
three may be the most stressful day for dogs housed in the
shelter because they have not yet begun to acclimate, are
responding to consistently unpredictable surroundings and
potentially unlike day 2, have reached their tolerance level
for novel, irritating stimuli. On day 3, the contact animals had
much lower cortisol levels than the no-contact animals (0.252
vs. 0.409 Ag/dl).
The stress response trend that has been seen in previous
research [6] was evident in the no-contact animals but not in
the contact animals. Previous research indicates that a dogs
Contact
No-contact
4 9
Day
540 C.L. Coppola et al. / Physiology & Behavior 87 (2006) 537 541
cortisol level begins to rise on day 2, peaks on day 3 4,
decreases until day 9 and then plateaus around day 9 10 when
housed in a shelter environment [6,7]. From the present data, it
was concluded that the contact animals were not experiencing
the spike in cortisol on day 3 typically observed in shelter dogs.
In the present study, animals in the no-contact group had
higher cortisol levels on day 3 than any other day. The contact
animals cortisol levels remained at approximately the same
level throughout the duration of the study (0.299, 0.252, 0.269
and 0.314 Ag/dl). The only difference between the two shelter
groups was that the contact animals participated in a human
contact session the 2nd day after their arrival at the shelter. This
one time session was influential in reducing the cortisol
response observed in shelter dogs. These results are in
agreement with a recent study [27] that examined the effect
of human interaction and diet on HPA axis activity of dogs in a
public animal shelter. In this study, it was concluded that
human interaction prevented sensitization of endocrine re-
sponse to novel situations.
For dogs, structuring a human contact session into shelter
arrival procedures shortly after entering the shelter system may
alleviate stress response. Studies [6,7] have recently shown that
interaction with humans may be an effective way to reduce the
cortisol stress response of shelter dogs. The results of the
present study are in accordance with both Hubrecht [28] and
Coppinger and Zuccotti [29] where it was concluded that social
isolation is the most stressful factor in a kennel environment
and that people are an important resource for dogs. Hubrecht
[28] also concluded that socialization by humans is especially
important for dogs that are housed singly.
In the shelter, socialization with other dogs and humans is
essential for good psychological well-being. Social isolation or
restriction is regarded as a major stressor for a social species
like the dog [30]. The withdrawal of human and/or conspecific
contact can be detrimental to mental health particularly to dogs
accustomed to such contact [31]. In fact, human contact may be
even more important than contact with another dog [30]. The
shelter environment does not normally promote positive social
interaction with humans; a controlled positive contact session
with a human can be influential in decreasing stress response
and facilitating acclimation to the facility and its stressors.
Additionally, less stress during the earlier portion of an
animals stay in the shelter may promote better animal welfare
overall.
Cortisol levels are affected by many variables and have a
high degree of individual variation. There were no indications
that human contact had any negative effects on the dogs and
positive human contact is generally regarded as pleasurable for
dogs as we showed in this study. Research in another
population of shelter dogs documented that a combination of
human interaction and high quality diet may have a positive
effect on behavior [32]. Stress may be further reduced in the
shelter environment by adding environmental enrichment (i.e.
toys, beds. companionship, food toys and complexity to the
enclosure), incorporating training into daily/weekly routine,
allowing for social interaction (human and conspecific) and
providing adequate exercise.
In addition to possibly reducing stress and improving
welfare, a human contact session can serve a dual purpose.
This session also allows the shelter to assess the dogs general
temperament and gain valuable information, which can be used
to improve the adoption process and facilitate successful
adoptions. There is a movement in the US for animal shelters
to at the very least preliminarily screen their dogs for
aggression while other shelters, if resources allow, evaluate
their dogs for temperament/personality/behavior in an effort to
provide more information and make a more appropriate match
between dog and owner.
Acknowledgements
The authors thank the Weld County Humane Society for
their participation and cooperation. We are also grateful for the
technical assistance provided by Mr. Robert Yemm and Dr.
Patrick Burns. Grandin Livestock Inc., Fort Collins, Colorado
and the Eugene V. and Clare E. Thaw Charitable Trust, Santa
Fe, New Mexico provided financial support for this study.
References
[1] Mason JW. A review of psychoendocrine research on the pituitary
adrenal cortical system. Psychosom Med 1968;30:576 607.
[2] Friedman SB, Ader R. Adrenocortical response to novelty and noxious
stimulation. Neuroendocrinology 1967;2:209 12.
[3] Hennessy MB. Hypothalamic pituitary adrenal responses to brief social
separation. Neurosci Biobehav Rev 1997:11 29.
[4] Muir JL, Pfister HP. Corticosterone and prolactin responses to
predictable and unpredictable novelty stress in rats. Physiol Behav
1986;37:285 8.
[5] Hanson JD, Larson ME, Snowden CT. The effects of control over high
intensity noise on plasma cortisol levels in rhesus monkeys. Behav Biol
1976;16:333 40.
[6] Hennessy MB, Davis HN, Williams MT, Mellott C, Douglas CW. Plasma
cortisol levels of dogs at a county animal shelter. Physiol Behav 1997;62:
485 90.
[7] 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.
[8] Dess NK, Linwick D, Patterson J, Overmier JB. Immediate and proactive
effects of controllability and predictability on plasma cortisol responses to
shocks in dogs. Behav Neurosci 1983;97:1005 116.
[9] Lynch JJ, McCarthy JF. The effect of petting on a classically conditioned
emotional response. Behav Res Ther 1967;5:55 62.
[10] Kirschbaum C, Hellhammer DH. Salivary cortisol in psychobiological
research: an overview. Neuropsychobiology 1989;22:150 69.
[11] Bassett JR, Marshall PM, Spillane R. The physiological measurement of
acute stress (public speaking) in bank employees. Int J Psychobiol 1987;5:
265 73.
[12] Blackshaw JK, Blackshaw AW. Limitation of salivary and blood cortisol
determination in pigs. Vet Res Commun 1989;13:265 71.
[13] Parrott RF, Misson BH. Changes in pig salivary cortisol in response to
transport stimulation, food and water deprivation, and mixing. Br Vet J
1989;145:501 5.
[14] Beerda B, Schilder MBH, Janssen NSCRM, Mol AJ. The use of saliva
cortisol, urinary cortisol and catecholamine measurements for a noninva-
sive assessment of stress response in dogs. Horm Behav 1996;30:272 9.
[15] Cook NJ, Schaefer AL, Lepage P, Jones SM. Salivary vs. serum cortisol
for the assessment of adrenal activity in swine. Can J Anim Sci 1996;76:
329 35.
 C.L. Coppola et al. / Physiology & Behavior 87 (2006) 537 541
[16] Groschl M, Biskupek-Sigwart J, Rauh M, Dorr HG. Measurement of
cortisol in saliva using a commercial radioimmunoassay developed for
serum. J Lab Med 2000;24:314 9.
[17] Cook NJ, Schaefer AL, Lepage P, Morgan Jones SD. Radioimmunoassay
for cortisol in pig saliva and serum. J Agric Food Chem 1997;45:395 9.
[18] 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.
[19] Lester SJ, Bellamy JEC, MacWilliams PS, Feldman EC. A rapid
radioimmunoassay method for the evaluation of plasma cortisol levels
and adrenal function in the dog. J Am Anim Hosp Assoc 1981;17:121 8.
[20] Reimers TJ, Mummery LK, McCann JP, Cowan RG, Concannon PW.
Effects of reproductive state on concentrations of thyroxine,3,5,3V-
triiodothyronine and cortisol in serum of dogs. Biol Reprod 1984;31:
148 54.
[21] Reimers TJ, Lawler DF, Sutaria PM, Correa MT, Erb HN. Effects of age,
sex, and body size on serum concentrations of thyroid and adrenocortical
hormones in dogs. Am J Vet Res 1990;51:454 7.
[22] Garnier F, Benoit E, Virat M, Ochoa R, Delatour P. Adrenal cortical
response in clinically normal dogs before and after adaptation to a housing
environment. Lab Anim 1990;24:40 3.
[23] Walker RF, Joyce BG, Dyas J. Salivary cortisol: 1. Monitoring changes in
normal adrenal activity. Immunoassays of steroids in saliva. Cardiff
Alpha Omega; 1984. p. 308 16.
[24] Tunn S, Mollmann H, Barth J, Derendorf H, Krieg M. Simultaneous
measurement of cortisol in serum and saliva after different forms of
cortisol administration. Clin Chem 1992;38:1491 4.
All in-text references underlined in blue are linked to publications on ResearchGate, letting you access and read them immediately.
541
[25] Riley V, Fitzmaurice MA, Spackman DH. Psychoneuroimmunologic
factors in neoplasia: studies in animals. Psychoneuroimmunology. New
York Academic Press; 1981. p. 31 102.
[26] Tuber DS, Miller DD, Caris KA, Halter R, Linden F, Hennessy MB. Dogs
in animal shelters: problems, suggestions and needed expertise. Psychol
Sci 1999;10:379 86.
[27] Hennessy MB, Voith VL, Hawke JL, Young TL, Centrone J, McDowell
AL, Linden F, Davenport GM. Effects of a program of human interaction
and alterations in diet composition on activity of the hypothalamic
pituitary adrenal axis in dogs housed in a public animal shelter. J Am Vet
Med Assoc 2002;221:65 71.
[28] Hubrecht RC. Dog housing and welfare. Wheathampstead. Herts UK
Universities Federation for Animal Welfare; 1993 Report no. 6.
[29] Coppinger R, Zuccotti J. Kennel enrichment: exercise and socialization of
dogs. J Appl Anim Welf Sci 1999;2:281 96.
[30] Wolfe TL. Policy, program and people: the three Ps to well-being.
Scientists center for animal welfare. MD Canine Research Environment
Bethesda; 1990. p. 41 7.
[31] Fox MW. Laboratory animal husbandry. Albany, NY State University of
New York Press; 1986.
[32] 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;5:253 73.