Content uploaded by Eva Hydbring Sandberg
Author content
All content in this area was uploaded by Eva Hydbring Sandberg
Content may be subject to copyright.
Short-Term Interaction
between Dogs and Their
Owners: Effects on Oxytocin,
Cortisol, Insulin and Heart
Rate—An Exploratory Study
Linda Handlin*†, Eva Hydbring-Sandberg‡,
Anne Nilsson†, Mikael Ejdebäck*, Anna Jansson§and
Kerstin Uvnäs-Moberg*†
*Systems Biology Research Centre, University of Skövde, Sweden
†Department of Animal Environment and Health, Swedish University
of Agricultural Sciences, Skara, Sweden
‡Department of Anatomy, Physiology, and Biochemistry, Swedish
University of Agricultural Sciences, Uppsala, Sweden
§Department of Animal Nutrition and Management, Swedish Univer-
sity of Agricultural Sciences, Uppsala, Sweden
ABSTRACT The aim of this exploratory study was to determine heart rate
and the levels of oxytocin, cortisol, and insulin in dogs and their owners in
response to a short-term interaction. In addition, the dogs’ behavior was
studied. The owners’ responses were compared with those obtained from a
control group. Ten female volunteers and their own male Labrador dogs par-
ticipated in an experiment during which the owner stroked, petted, and talked
with her dog during the first 3 minutes. Blood samples were collected from
both dog and owner before (0) and at 1, 3, 5, 15, 30, and 60 minutes after the
start of the interaction. Blood samples were analyzed by EIA. Heart rate was
monitored telemetrically. The data were analyzed using linear mixed models
and paired t-tests. The dogs’ oxytocin levels were significantly increased 3
minutes after the start of the interaction (p= 0.027). Cortisol levels were
significantly increased after 15 and 30 minutes (p= 0.004 and p= 0.022,
respectively), and heart rate was significantly decreased after 55 minutes
(p= 0.008). The dogs displayed normal behaviors during the experiment. The
owners’ oxytocin levels peaked between 1 and 5 minutes after interaction
(p = 0.026). No such effect was seen in the controls. Cortisol levels displayed
a significant decrease at 15 or 30 minutes in both owners and controls, and
insulin levels did so at 60 minutes (p= 0.030, p= 0.002 and p= 0.002,
301 Anthrozoös DOI: 10.2752/175303711X13045914865385
ANTHROZOÖS VOLUME 24, ISSUE 3 REPRINTS AVAILABLE PHOTOCOPYING © ISAZ 2011
PP. 301–315 DIRECTLY FROM PERMITTED PRINTED IN THE UK
THE PUBLISHERS BY LICENSE ONLY
Address for correspondence:
Linda Handlin,
Systems Biology Research
Centre,
University of Skövde,
Box 408, SE-541 28
Skövde, Sweden.
E-mail: linda.handlin@his.se
AZ VOL. 24 (3).qxp:Layout 1 6/29/11 3:35 PM Page 301
302 Anthrozoös
Short-Term Interaction between Dogs and Their Owners…
p< 0.0001, respectively). Heart rate decreased significantly in the owners at 55 and 60 minutes
(p= 0.0008) but not in the controls. In conclusion, short-term sensory interaction between dogs and
their owners influences hormonal levels and heart rate. However, further studies need to be
performed in order to better understand the effects of interaction between dogs and their owners.
Keywords: cortisol, heart rate, human–dog interaction, insulin, oxytocin
Human–animal interaction (HAI) has been shown to have positive effects on health
and well-being in humans. The acquisition of pets can result in a reduction in health
problems and an improvement in perceived health (Serpell 1991). Pet owners have
been shown to have lower levels of risk factors for cardiovascular disease, and after acute
myocardial infarction, dog owners are significantly less likely to die within 1 year, compared with
those who do not own dogs (Friedmann and Thomas 1995). Owning a pet is associated with
lower heart rate and blood pressure during basal and stressed conditions (Allen, Shykoff and
Izzo 2001; Allen, Blascovich and Mendes 2002). In addition, anxiety decreases in the presence
of a dog (Wilson 1991) and children having a dog present in their classroom display increased
social competence (Hergovich et al. 2002; Kotrschal and Ortbauer 2003).
The positive health consequences associated with HAI may be caused by oxytocin re-
lease induced by positive emotions such as affection and love (Uvnäs-Moberg 1997; 1998) and
by the physical interaction that takes place between the human and animal. The physical
interaction between humans and dogs involves various types of non-noxious sensory stimu-
lation such as touch, light pressure, warmth, and stroking as well as olfactory, auditory, and
visual cues.
Non-noxious sensory stimulation gives rise to physiological effects in anesthetized rats; for
example, decreased activity in the hypothalamic pituitary adrenal (HPA)-axis and in the sym-
patho-adrenal system, resulting in decreased cortisol and adrenalin levels and lowered blood
pressure. It further increases oxytocin levels and influences the levels of gastrointestinal hor-
mones, as a consequence of efferent vagal nerve activation (Kurosawa et al. 1982; Araki et al.
1984; Stock and Uvnäs-Moberg 1988; Uvnäs-Moberg et al. 1992; Kurosawa et al. 1995). In
unanesthetized rats, both physiological and behavioral effects are induced by non-noxious
sensory stimulation. Stroking of the abdomen (40 strokes/minute for 5 minutes) decreases
pulse rate and blood pressure for several hours (Lund et al. 1999), pain thresholds are in-
creased, and a sedative effect is induced. In addition, oxytocin is released (Agren et al. 1995;
Uvnäs-Moberg et al. 1996). Repeated exposure to stroking gives rise to long-lasting effects
such as an increased pain threshold, decreased blood pressure, and decreased levels of gas-
trointestinal hormones and energy expenditure, resulting in weight gain (Holst, Uvnäs-Moberg
and Petersson 2002; Lund et al. 2002; Holst et al. 2005). Newborn rats subjected to large
amounts of non-noxious sensory stimulation in the form of maternal sensory interaction dis-
play reduced fear, increased social interaction, and increased function of oxytocin receptors
in the amygdala as adults (Liu et al. 1997; Francis et al. 2002).
Similar effects, including decreased cortisol levels, can be observed in humans in response
to non-noxious sensory stimulation such as massage, skin-to-skin contact between mothers
and infants, and suckling in breastfeeding mothers (Uvnäs-Moberg et al. 1990; Nissen et al.
1996; Uvnäs-Moberg 1996; Uvnäs-Moberg and Eriksson 1996; Handlin et al. 2009). Taken to-
gether, these data show that non-noxious sensory stimulation has stress-reducing effects by
reducing the activity in the HPA-axis and by decreasing and increasing the activity in certain
aspects of the sympathetic and parasympathetic/vagal nervous systems, respectively.
❖
AZ VOL. 24 (3).qxp:Layout 1 6/29/11 3:35 PM Page 302
Oxytocin may mediate some of the effects mentioned above by actions in the brain, in
particular since some of the effects induced by non-noxious sensory stimulation are reversed
following the administration of oxytocin antagonists (Uvnäs-Moberg 1998; Uvnäs-Moberg and
Petersson in press). Oxytocin is produced in the supraoptic nucleus (SON) and paraventricu-
lar nucleus (PVN) of the hypothalamus and was originally described as a hormone released into
the circulation during labor and suckling (Richard, Moos and Freund-Mercier 1991). However,
oxytocin is also released into important regulatory areas in the brain from nerves originating in
the PVN. Oxytocin stimulates social interactive behavior and promotes attachment between
individuals (Carter 1998). Oxytocin also induces, for example, anxiolytic-like effects and seda-
tive effects (Uvnäs-Moberg et al. 1994; Amico et al. 2004), decreases cortisol levels and blood
pressure, and influences the release of gastrointestinal hormones, for example, insulin (Pe-
tersson et al. 1996; Petersson, Hulting and Uvnäs-Moberg 1999; Petersson et al. 1999; Holst,
Uvnäs-Moberg and Petersson 2002).
Since non-noxious sensory stimulation gives rise to a multitude of effects that may in part
be mediated by oxytocin in both humans and animals, it is likely that oxytocin release and
oxytocin-mediated effects are induced during interaction between humans and dogs. Such ef-
fects might explain the health-promoting effects of HAI. This idea is supported by previous
studies which show that oxytocin is released in both dogs and humans when they interact
physically (Odendaal and Meintjes 2003; Miller et al. 2009).
The aim of this exploratory study was to test the hypothesis that oxytocin release and oxy-
tocin-mediated effects are induced in both dogs and their owners during a short period of in-
teraction characterized by caressing and stroking. To test the hypothesis, oxytocin levels were
measured in blood samples collected before, during, and after a short-term interaction be-
tween dogs and their owners. Since oxytocin influences the activity of the HPA axis and the
autonomic nervous system, cortisol levels were measured to reflect the activity in the HPA
axis, and insulin levels were measured to reflect vagal nerve tone. Heart rate was measured
to reflect both sympathetic and parasympathetic activity. The dogs’ behaviors were analyzed
to check that the dogs were well and were not stressed by the experiment.
Methods
Participants
Ten privately owned male Labrador dogs and their female owners were recruited through in-
formation provided at local workplaces and local veterinary clinics. The owners were informed
about the aim of the study and about the experimental setup. Women older than 30 years who
owned a male Labrador older than 1 year were included in the study. Owners and their dogs
who participated in the study did not have any illnesses. Ten female volunteers who did not
own a dog served as controls. The mean age of the owners and controls was 53 years
(SD = 10), and 42 years (SD = 8), respectively, and the mean age of the dogs was 4.7 years
(SD = 2.6). There was no significant difference in age between owners and controls. Due to
ethical and practical reasons, it was not possible to perform control experiments on the
privately owned dogs.
The study was conducted at the Swedish University of Agricultural Sciences in Skara,
Sweden. The experimental procedure for the humans was approved by the Local Ethics Com-
mittee in Uppsala, and the procedure for the dogs was approved by the Animal Ethics Com-
mittee in Uppsala. The use of privately owned dogs was approved by the National Board of
Agriculture. Before the experiment started, the owners were once more informed about the
303 Anthrozoös
Handlin et al.
AZ VOL. 24 (3).qxp:Layout 1 6/29/11 3:35 PM Page 303
study, were given the opportunity to ask any questions regarding the experiment, and were
made aware that they could end their participation at any time. The owner then signed a written
consent form for participation in the study for both the dog and herself.
Experimental Setting
The owner and her dog arrived together at the testing facility, which consisted of a plain room
containing a desk, four chairs, a bookcase, and a water bowl for the dog. The owners sat in
a chair during the entire experiment. In addition, four other people were present in the room
during the experiment: an animal caretaker, a nurse, one person taking care of the blood
samples, and one person videotaping the experiment. None of these four persons were in
contact with the dog or the owner during the experiment except for the animal caretaker and
the nurse during the insertion of catheters and sampling of blood.
Preparations
An indwelling catheter was inserted into the cubital vein of the dog owners and the controls.
In order to facilitate the insertion of catheters and sampling of blood, the dogs were shaved
on the dorsal side of the distal part of a forelimb, where a local anesthetic plaster with prilo-
caine and lidocaine (EMLA®, AstraZeneca) was attached. The plaster was applied for 45 min-
utes and then an intravenous catheter was inserted into the cephalic vein. The catheter was
covered with Vetrap (CM), in order to prevent the dog from licking it.
The dogs were also shaved on a small area on the lateral side of the chest around which
a heart rate monitor (s610i TM, Polar precision performance, Polar Electro) was attached, with
the receiver placed a short distance away. The recording range was 30–240 beats/min and
the accuracy of the recordings was ± 1 beat/min. In order to maximize contact between the
heart rate monitor and the dogs’ skin, electrode gel (Blågel, Cefar) was used.
Heart rate monitors of the same type as for the dogs were attached around the chests of
the dog owners and the controls, with the receivers placed around their wrists. Heart rate was
monitored every 15th second in both dogs and humans. Due to technical problems, heart
rate was only measured in five controls.
The Interaction Experiment
Before the experiment started, the owner sat in a chair with her dog unrestrained, sitting or lying
beside her. The owner approached her dog at time point zero and started to pet and stroke
different parts of the dog’s body and talked to him for 3 minutes. The owner was then in-
structed to remain sitting in her chair and not to touch her dog for the rest of the experiment,
which lasted for a total of 60 minutes. If the dog attempted to interact with the owner during
the remaining time of the experiment, the owner was instructed to ignore this, with the result
that the dog stopped its attempts almost immediately. Verbal communication was allowed
during the whole experiment.
The conditions for the control group were the same as for the owners, with the exception
that there was no dog present: the participants went through the same preparations, sat in the
same chair in the same room with the same people present (except for the animal caretaker),
and blood samples were drawn in the same way at the same time points as for the owners.
The goal was to perform all experiments, for the owners and the controls, during the
evening, but due to the participants’ work schedules, some of the experiments were performed
during the morning (4 owners and 5 controls).
304 Anthrozoös
Short-Term Interaction between Dogs and Their Owners…
AZ VOL. 24 (3).qxp:Layout 1 6/29/11 3:35 PM Page 304
Blood Sampling
The first blood samples were taken simultaneously from both the dog and the owner 30 min-
utes after insertion of the catheter and just before the owner started to interact with her dog
(basal = 0 min). Blood samples were then taken from both the dog and the owner at 1, 3, 5,
15, 30, and 60 minutes after the start of the interaction. Insertion of catheters and sampling
of blood in the dogs and humans were performed by the same experienced animal caretaker
and experienced nurse, respectively.
The blood samples were collected in 4 ml EDTA tubes containing 0.2 ml aprotinin
(Trasylol®, Bayer AB). The samples were immediately put on ice and then centrifuged at
1600 ⫻g for 20 minutes at +4ºC after which the plasma was collected and stored at
–20ºC, until analysis.
Hormone Analysis
Oxytocin levels were determined in humans and dogs using Correlate-EIATM Oxytocin En-
zyme Immunoassay Kit according to the manufacturer’s instructions (Assay Designs, Inc.
Ann Arbor, USA) (sensitivity 11.7 pg/ml, precision 9.1%). Cortisol levels were determined
using the DSL-10-2000 ACTIVE®Cortisol Enzyme Immunoassay Kit according to the man-
ufacturer’s instructions (Diagnostic Systems Laboratories, Inc. Texas, USA) (sensitivity 2.76
nmol/l, precision 10.3%). The dogs’ insulin levels were determined using the Mercodia Ca-
nine Insulin ELISA 10-1203-1 according to the manufacturer’s instructions (Mercodia AB,
Uppsala, Sweden) (sensitivity 0.01g/l, precision 4.6%), and the humans’ insulin levels
were determined using the Mercodia Insulin ELISA 10-1113-10 according to the manu-
facturer’s instructions (Mercodia AB, Uppsala, Sweden) (sensitivity 1 mU/l, precision 5%).
Standards and controls were always included, as recommended by the manufacturers.
Before the oxytocin analysis, the samples from the humans were diluted five times in the
assay buffer. Before the cortisol analysis, the samples from the dogs were diluted two times
in zero standard buffer.
All washing procedures were performed using an Anthos Fluido microplate washer (Anthos
Labtec Instruments GmbH), and the absorbance was read using a Multiskan Ex microplate
photometer (Thermo Electron Corporation). The color development of the samples was read
at 405 nm for oxytocin and at 450 nm for cortisol and insulin, with background correction at
580 nm for oxytocin and 620 nm for cortisol. Ascent software was used for creation of
standard curves, curve fitting, and calculation of concentrations (Ascent software ver 2.6 for
iEMS Reader MF and multiscan).
One dog was excluded from the oxytocin analysis because his hormone levels were out-
side the range of detection.
Observations of the Dogs’ Behaviors
The entire interaction experiment (60 min) was videotaped, in order to control for how the dogs
behaved and experienced the situation. The total time the dog spent sitting, standing, lying
down (i.e., both on the chest and on the side), or walking around was recorded. The total time
the dog was resting (i.e., put his head on the floor) while lying down, the number of times the
dog changed the position of his head while lying down (i.e., when the dog moved his head from
left to right or from right to left), and the number of times the dog changed body positions
while lying down (i.e., how many times the dog moved his whole body from one side to the
other) were recorded. The number of times the dog licked himself around the mouth was also
recorded, since this may be an indicator of stress in dogs.
305 Anthrozoös
Handlin et al.
AZ VOL. 24 (3).qxp:Layout 1 6/29/11 3:35 PM Page 305
Statistical Analysis
The data were analyzed using SAS version 9.1 for Windows (Cary, NC, USA; 2002) and Sta-
tistical Package for the Social Sciences (SPSS/PASW) version 17.0 (Chicago, IL, USA; 2009).
Mean values with corresponding standard errors (SE) were used to describe the hormone lev-
els and the heart rate of the dogs, owners, and controls, as well as to describe the different
behaviors studied in the dogs.
The distributions of hormone levels for dogs and humans were positively skewed and
therefore normalized by logarithmic transformation (log10) before statistical analysis was per-
formed. The heart rate monitors registered the heart rate every 15th second, but only the
recordings obtained at each 5th min were used in the statistical analysis.
Sampling time was considered as a categorical predictor and the change in heart rate and hor-
mone levels at specific time points compared with the start of the dog–owner interaction (0 min)
was analyzed using linear mixed models in the MIXED procedure of SAS, one model for each trait.
For dogs, the models contained sampling time, and the model of cortisol also contained a pre-
dictor representing time of day for blood sampling (morning or evening); timing had no significant
effect on the other traits and was not included in these models. For humans, the models contained
sampling time, a variable representing group (owner or control), and the interaction between time
and group; again, the cortisol model contained a predictor for time of day. Correlation between
samples within participants was accounted for by a REPEATED statement and, due to the rela-
tively large number of unequally distributed time points for hormones, a compound symmetry
correlation structure was applied. Satterthwaite denominator degrees of freedom were specified.
The oxytocin level at 3 min was thus compared with the 0 min level. In contrast, the corti-
sol levels at 15 and 30 min were compared with 0 min. Likewise, the insulin level at 60 min and
heart rate at 55 and 60 min were tested. Time points for comparisons were selected based
on previous research and experience, and predicted values (least-squares means) were
compared using t-tests. P values < 0.05 were considered statistically significant. In dogs, a total
of 63, 70, 70, and 129 observations were used in the models of oxytocin, cortisol, insulin, and
heart rate, respectively; in humans, 139, 139, 137, and 195 observations were used,
respectively. The models were validated by examining the normality of raw residuals.
In a complementary analysis, extreme values of oxytocin and cortisol were considered.
The maximum oxytocin level obtained at 1, 3, or 5 min was recorded for each participant. In
the same way, the maximum (in dogs) and minimum (in humans) cortisol levels obtained at 15
or 30 minutes were recorded. Again, these time points were chosen based on earlier scientific
evidence. Paired t-tests were calculated to test for differences between the extreme values and
the basal values.
Results
Hormone levels and heart rate for all participants are presented in Tables 1 and 2. For
information about differences between least-squares means at selected time points, standard
errors, and p values generated in the linear mixed models, see Table 3.
Dogs
Hormone Levels: The dogs’ oxytocin levels were significantly increased 3 min after the start of
the dog–owner interaction (p= 0.027) (Table 3). In addition, the dogs’ peak oxytocin levels
recorded at 1, 3, or 5 min were significantly higher compared with the levels collected at time
point 0 min (t= 2.99; df = 8; p= 0.017) (Figure 1; Table 1).
306 Anthrozoös
Short-Term Interaction between Dogs and Their Owners…
AZ VOL. 24 (3).qxp:Layout 1 6/29/11 3:35 PM Page 306
The dogs’ mean cortisol levels exhibited a delayed and protracted rise, and cortisol levels
were significantly increased after 15 and 30 min after the start of the dog–owner interaction,
when compared with levels obtained before the interaction started (p= 0.004 and p= 0.022,
respectively) (Table 3). A significant peak in the dogs’ cortisol levels during the same time period
could not be demonstrated with the paired t-test (p= 0.146) (Figure 2; Table 1). The levels of
cortisol were 41% higher in the morning than in the evening (p= 0.048).
Insulin levels did not change significantly during the experiment (Figure 3; Tables 1 and 3).
Heart Rate: There was a significant decrease in the dogs’ heart rate at 55 min compared with
at the start of the interaction (p= 0.008). In contrast, there was no significant change at 60 min
(Figure 4; Tables 2 and 3).
Behavioral Observations: These were made from the videotapes and the results are summa-
rized in Table 4. The dogs displayed normal behaviors during the experiment. They walked
around for a short while but were mostly resting.
307 Anthrozoös
Handlin et al.
Table 1. Hormone concentrations during the interaction experiment, and maximum levels of
oxytocin and maximum levels (in case for the dogs) and minimum levels (in case of the
humans) of cortisol for the participants (10 male Labrador dogs, 10 female owners, and 10
control persons). Means (SE) are shown and are based on non-transformed data.
0 min 1 min 3 min 5 min 15 min 30 min 60 min Min/Max
Values
Oxytocin Levels (pmol/l)
Dogs 155.8 211.2 236.9 178.6 163.5 157.5 157.5 251.8
(26.9) (30.7) (38.7) (29.6) (34.5) (36.0) (41.1) (34.5) (max)
Owners 168.5 169.8 180.6 170.2 146.4 171.3 165.1 187.0
(34.6) (34.1) (34.4) (27.8) (34.7) (34.2) (26.3) (33.6) (max)
Controls 208.6 208.2 212.1 212.5 215.6 198.0 212.4 166.1
(62.0) (64.3) (68.2) (70.0) (68.0) (62.9) (67.4) (43.7) (max)
Cortisol Levels (nmol/l)
Dogs 168.4 169.4 168.1 180.1 224.1 202.8 190.2 237.3
(14.8) (16.1) (15.3) (17.8) (32.5) (18.3) (18.8) (30.3) (max)
Owners 389.8 382.7 382.7 387.6 362.1 331.6 305.2 316.9
(119.7) (107.4) (109.9) (119.6) (107.9) (80.1) (62.6) (80.5) (min)
Controls 381.5 379.4 375.5 371.9 352.7 332.6 345.5 319.9
(30.9) (25.9) (27.8) (28.9) (24.2) (33.5) (37.2) (26.5) (min)
Insulin Levels (pmol/l)
Dogs 37.5 32.6 28.5 30.8 34.6 32.4 42.9 –
(7.8) (5.4) (3.4) (3.8) (8.9) (6.7) (8.6)
Owners 159.6 149.0 142.5 151.8 142.2 126.8 101.4 –
(42.5) (36.8) (31.7) (38.8) (53.9) (44.4) (35.2)
Controls 155.2 159.4 154.7 148.1 159.0 114.1 51.4 –
(67.6) (68.4) (58.4) (48.7) (44.9) (42.3) (32.7)
AZ VOL. 24 (3).qxp:Layout 1 6/29/11 3:35 PM Page 307
Dog Owners and Controls
Hormone Levels: Neither owners nor controls showed a significant change in oxytocin levels
after 3 min of dog–owner interaction (Table 3). However, the owners’ peak oxytocin levels
recorded at 1, 3, or 5 min were significantly higher compared with the levels collected at time
point 0 min (t= 2.66; df = 9; p= 0.026). Such an effect was not seen in the controls (p=
0.417) (Figure1; Tables 1 and 3).
308 Anthrozoös
Short-Term Interaction between Dogs and Their Owners…
Table 2. Heart rate (beats/min) during the interaction experiment for the participants (10 male
Labrador dogs, 10 female owners, and 10 control persons). Mean (SE) values are shown and
are based on non-transformed data.
0 51015202530354045505560
min min min min min min min min min min min min min
Dogs 94 94 83 88 84 67 75 88 87 82 80 67 76
(11) (8) (9) (9) (9) (7) (3) (10) (14) (15) (15) (8) (4, 2)
Owners 78 78 78 76 74 74 74 71 72 71 72 71 71
(5) (5) (4) (4) (4) (4) (4) (4) (3) (3) (4) (4) (4)
Controls 74 68 70 70 72 68 70 72 73 70 73 71 71
(6) (5) (5) (5) (6) (6) (5) (6) (6) (5) (3) (6) (4)
Table 3. Back-transformed least-squares means (LSM) at start of dog–owner interaction
(0 min) and at selected time points, standard errors (SE), and pvalues generated from the
linear mixed models of hormone levels and heart rate.
Trait and Time Point LSM SE p
Dogs
Oxytocin at 0 and 3 min (pmol/l) 139.5; 210.1 1.20 0.027
Cortisol at 0 and 15 min (nmol/l) 162.2; 205.1 1.08 0.004
Cortisol at 0 and 30 min (nmol/l) 162.2; 195.0 1.08 0.022
Insulin at 0 and 60 min (pmol/l) 32.4; 31.4 1.22 0.882
Heart rate at 0 and 55 min (beats/min) 94.2; 66.9 10.02 0.008
Heart rate at 0 and 60 min (beats/min) 94.2; 83.1 10.32 0.283
Owners
Oxytocin at 0 and 3 min (pmol/l) 147.2; 160.3 1.07 0.191
Cortisol at 0 and 15 min (nmol/l) 297; 263.5 1.06 0.055
Cortisol at 0 and 30 min (nmol/l) 297; 271.6 1.06 0.136
Insulin at 0 and 60 min (pmol/l) 129.5; 69.8 1.21 0.002
Heart rate at 0 and 55 min (beats/min) 78.1; 71.4 1.96 0.0008
Heart rate at 0 and 60 min (beats/min) 78.1; 71.4 1.96 0.0008
Controls
Oxytocin at 0 and 3 min (pmol/l) 159.2; 155.2 1.07 0.697
Cortisol at 0 and 15 min (nmol/l) 369.0; 345.1 1.06 0.266
Cortisol at 0 and 30 min (nmol/l) 369.0; 318.4 1.06 0.015
Insulin at 0 and 60 min (pmol/l) 84.4; 28.4 1.23 < 0.0001
Heart rate at 0 and 55 min (beats/min) 73.8; 71.2 2.77 0.350
Heart rate at 0 and 60 min (beats/min) 73.8; 71.0 2.77 0.314
AZ VOL. 24 (3).qxp:Layout 1 6/29/11 3:35 PM Page 308
The owners’ cortisol levels tended to be decreased at 15 min after the start of the inter-
action with the dog (p= 0.055) but not at 30 min (p= 0.14) (Table 3). Their minimum cortisol
309 Anthrozoös
Handlin et al.
220
195
170
145
120
0
Oxytocin (pmol/l)
Dogs
10 20 30 40 50 60
Time (min)
Owners
Controls
Figure 1. Predicted levels of oxytocin (pmol/l) in 10 male Labrador dogs,
10 female owners, and 10 female control persons; back-transformed
least-squares means from a linear mixed model. The first blood sample
was taken immediately before the owner started to interact with her dog
(0 min) and other samples were taken 1, 3, 5, 15, 30, and 60 min later.
Standard errors for oxytocin levels at 0 and 3 minutes in dogs = 1.2, in
owners = 1.22, and in controls = 1.22.
400
350
300
250
200
150
100
0
Cortisol (nmol/l)
Dogs
10 20 30 40 50 60
Time (min)
Owners
Controls
Figure 2. Predicted levels of cortisol (nmol/l) in 10 male Labrador dogs,
10 female owners, and 10 female control persons; back-transformed
least-squares means from a linear mixed model. The first blood sample
was taken immediately before the owner started to interact with her dog
(0 min) and other samples were taken 1, 3, 5, 15, 30, and 60 min later.
Standard errors for cortisol levels at 0, 15 and 30 minutes in dogs = 1.1,
in owners = 1.18, and in controls = 1.18.
AZ VOL. 24 (3).qxp:Layout 1 6/29/11 3:35 PM Page 309
310 Anthrozoös
Short-Term Interaction between Dogs and Their Owners…
140
120
100
80
60
40
20
0
Insulin (pmol/l)
Dogs
10 20 30 40 50 60
Time (min)
Owners
Controls
Figure 3. Predicted levels of insulin (pmol/l) in 10 male Labrador dogs,
10 female owners, and 10 female control persons; back-transformed
least-squares means from a linear mixed model. The first blood sample
was taken immediately before the owner started to interact with her dog
(0 min) and other samples were taken 1, 3, 5, 15, 30, and 60 min later.
Standard errors for insulin levels at 0 and 60 minutes in dogs = 1.22, in
owners = 1.03, and in controls = 1.03.
100
95
90
85
80
75
70
65
60
0
Heart Rate (beats/min)
Dogs
10 20 30 40 50 60
Time (min)
Owners
Controls
Figure 4. Predicted levels of heart rate (beats/min) in 10 male Labrador
dogs, 10 female owners, and 10 female control persons; least-squares
means from a linear mixed model. Standard errors at 0, 55, and 60
minutes in dogs = 10.08, in owners = 4.00, and in controls = 5.66.
levels reached at 15 or 30 min were significantly decreased compared with the levels collected
at time point 0 min (t= –2.573; df = 9; p= 0.030) (Figure 2; Table 1).
The controls did not display a significant decrease in cortisol levels at 15 min (p= 0.266),
but did so at 30 min (p= 0.015) (Table 1). In addition, their minimum cortisol levels recorded
at 15 or 30 min were significantly decreased compared with the levels collected at time point
AZ VOL. 24 (3).qxp:Layout 1 6/29/11 3:35 PM Page 310
0 min (t= –4.275; df = 9; p= 0.002) (Figure 2; Table 1). The levels of cortisol for both owners
and controls were 80% higher in the morning than in the evening (p= 0.010).
In both owners and controls, there was a significant decrease in insulin levels at 60 min
(p= 0.0018 and p< 0.001, respectively) (Figure 3; Table 3).
Heart Rate: Heart rate was significantly decreased at 55 and 60 min in the owners (p= 0.0008
and p= 0.0008, respectively). In contrast, no change in heart rate was seen in the controls
(Figure 4; Table 3).
Differences between Owners and Controls Over Time
There was a significant statistical interaction between time and group for insulin and heart
rate (p= 0.045 and p= 0.011, respectively) but not for oxytocin and cortisol (p= 0.650 and
p= 0.906, respectively).
Discussion
The aim of this exploratory study was to test the hypothesis that oxytocin release and oxytocin-
mediated effects are induced in both dogs and their owners in response to physical interac-
tion. The results show that short-term interaction between a dog and its owner is associated
with a significant increase in oxytocin and cortisol levels in the dog. In addition, oxytocin in-
creased significantly in the owners but not in the controls, cortisol and insulin levels decreased
in both the owners and the controls, while heart rate decreased significantly only in the own-
ers. To our knowledge, the present study is unique since it was performed under standard-
ized conditions with an equal focus on the humans and dogs, and repeated observations
were made before, during, and after interaction between the dog and its owner.
We chose to study male Labrador dogs and their female owners in order to keep variation
due to breed and sex steroid levels to a minimum. In addition, Labradors are one of the most
common companion dogs and are friendly and prone to interaction with humans, which is ad-
vantageous when studying human–animal interaction
In a previous study performed by Odendaal and Meintjes in 2003, some effects were noted
in response to interaction between dogs and humans. The levels of -endorphin, oxytocin,
prolactin, phenyl acetic acid, and dopamine all increased in the dogs and humans after inter-
action, whereas cortisol increased in the dogs and decreased in the humans (Odendaal and
Meintjes 2003). Although the results from that study display effects of interaction, the differ-
ences between it and our study regarding participants, experimental layout (only two blood
311 Anthrozoös
Handlin et al.
Table 4. Behavior of 10 male Labrador dogs during 60 min of human
interaction; means and their standard errors (SE).
Behavior Mean (SE)
Total time sitting (min:sec) 4:30 (1:24)
Total time standing (min:sec) 2:00 (0:42)
Total time lying down (min:sec) 52:00 (2:00)
Total time walking around (min:sec) 2:00 (0:36)
Total time resting while lying down (min:sec) 30:00 (4:14)
Licking around mouth (number of times) 44 (13)
Position changes of the head while lying down (number of times) 27 (4)
Position changes total (number of times) 17 (4)
AZ VOL. 24 (3).qxp:Layout 1 6/29/11 3:35 PM Page 311
samples collected before interaction and one collected between 4 and 24 minutes versus
repeated blood samples collected at different time points during the experiment), and differ-
ent methods for analyzing hormone levels (yielding extremely different hormone levels) makes
further comparisons difficult.
Recently, Miller et al. (2009) showed that women increase their oxytocin levels significantly
after interaction with their own dog. However, the same response was not observed in men.
That study focused only on humans, and only pre- and post observations were made (Miller
et al. 2009).
Nagasawa et al. (2009) reported increased oxytocin levels in the urine of humans in re-
sponse to gaze and touch (Nagasawa et al 2009). However, the relationship between oxytocin
levels in plasma, which mainly reflect oxytocin release from the pituitary, and oxytocin detected
in urine is at present unclear (Uvnäs-Moberg, Handlin and Petersson 2011).
Touch and massage-like stroking of rats and close physical contact in humans increases
oxytocin levels (Stock and Uvnäs-Moberg 1988; Nissen et al. 1995; Lund et al. 2002). Against
this background, it is likely that the distinct and short-lasting rise of oxytocin levels in the dogs
and the increased peak levels in the owners during our experiment were caused by the stroking
and petting performed by the owner.
In addition, non-noxious sensory stimulation induces stress-reducing effects in many dif-
ferent species. Rats being stroked on the abdomen show decreased blood pressure (Lund et
al. 1999). Cows being brushed on the abdomen show decreased heart rate and cortisol lev-
els (E. Wredle, personal communication). Skin-to-skin contact between mother and infant in-
duces lowering of blood pressure and cortisol levels in mothers and decreases cortisol levels
and increases cutaneous circulation in the infants (Nissen et al 1996; Uvnäs-Moberg 1996;
Uvnäs-Moberg and Eriksson 1996; Morelius, Theodorsson and Nelson 2005; Jonas et al.
2008; Handlin et al. 2009). Further, the levels of gastrointestinal hormones, including insulin,
are also influenced by stroking (Uvnäs-Moberg et al. 1992; Holst et al. 2005).
Since oxytocin is released by sensory stimulation, the interaction between the dogs and
owners may have decreased cortisol levels and heart rate via oxytocin released into the brain.
During the interaction experiment, though, the dogs displayed an increase in cortisol levels. A
rise in cortisol levels is often connected with high stress levels. It may, however, also reflect ini-
tiation of physical activity. In the present experiment, the dogs were behaviorally activated in
response to the interaction. Since cortisol levels in the circulation rise with a 15 to 20 min delay,
we suggest that the increase in the dogs’ cortisol levels reflect an increase in the locomotor
activity induced by the interaction with the owner. The increase was not confirmed by the com-
plementary paired t-test of maximum values at 15 or 30 min, which is probably explained by
the small number of participants.
The expected fall in insulin levels normally induced by sensory stimulation was not ob-
served in the present study. The owners were instructed not to feed their dogs just before ar-
riving at the testing facility. However, it turned out that some dogs had received food just before
arriving. A different pattern of insulin levels might have been obtained if all dogs had either
been fasted or fed before the start of the experiment, since feeding influences insulin levels.
The dogs’ heart rate was significantly decreased after 55 but not after 60 minutes. Oxytocin
induces, via effects in the brain, a decrease in sympathetic and an increase in parasympathetic
nervous tone. These changes influence cardiovascular function and hence oxytocin released
in the brain may have contributed to the temporary decrease in heart rate observed. The rea-
son for not seeing a significant decrease also after 60 minutes is probably due to the fact that
312 Anthrozoös
Short-Term Interaction between Dogs and Their Owners…
AZ VOL. 24 (3).qxp:Layout 1 6/29/11 3:35 PM Page 312
a blood sample was collected at this time point, which may have resulted in a slight activation
of the sympathetic nervous system.
It is clear that the response in dogs to dog–owner interaction is complex and involves en-
docrine and physiological reactions reflecting both activation and relaxation. In the present study,
the interaction experiment was performed in an unfamiliar room. However, the time that had
elapsed from entering the room until the start of the experiment gave the dogs time to adjust to
the unfamiliar surroundings. The results from the behavior analysis indicated that the dogs dis-
played normal behavior and therefore they seemed to tolerate the experimental situation well.
The owners’ oxytocin levels peaked between 1 and 5 min after interaction. This was not
seen in the controls. These data confirm the results of Miller et al. (2009) and Odendaal and
Meintjes (2003). The design of the present experiment, allowing the owner to be in the same
room as the dog as well as being aware of the interaction before it started, might explain why
the rise in circulating oxytocin levels occurred at different time points.
Both owners and controls displayed decreased cortisol levels over time. Perhaps the ex-
perimental situation, involving blood sampling, was perceived as stressful, which might have
caused a rise in cortisol levels in both groups, and then a decrease over time in both.
In addition, insulin levels fell over time in both owners and controls. Since feeding was not
controlled for, any effects caused by the sensory interaction in the owners might have been
concealed by parallel feeding-related changes in glucose and insulin levels.
The observation that the owners’ heart rate decreased significantly suggests that interaction
with the dog might have induced a slight anti-stress effect in the owners. This effect may be a
consequence of the oxytocin released in the brain caused by the sensory interaction. This find-
ing may be of great importance, since it may be easier to demonstrate a decreased activity in
the sympathetic nervous system than in the HPA-axis as a consequence of (sensory) interaction.
The ambition was to perform all experiments, for the owners and the controls, during the
evening, but due to the participants’ work schedules’ some of the experiments were per-
formed during the morning (4 owners and 5 controls). Oxytocin and insulin are not known to
have circadian rhythms, and hence the results for these hormones are not likely to have been
affected by the experiment being conducted at different times of day. In contrast, cortisol has
a circadian rhythm, with levels being highest in the morning and decreasing throughout the day
(Arlt and Stewart 2005). In line with this, cortisol levels were significantly higher in the morning
than in the afternoon in both humans and dogs.
Our results show that there is a release of oxytocin and that some oxytocin-mediated ef-
fects can be observed in dogs and their owners when they interact with each other. Since this
is an exploratory study with only 10 dog–owner pairs participating, the results need to be
interpreted with caution and further studies need to be performed with a larger number of par-
ticipants and under even more standardized conditions. This will facilitate better recording and
understanding of the physiological and behavioral responses in dogs and their owners as a
consequence of interaction. In addition, it would be interesting to study the consequences of
interaction between both female and male dogs and female and male owners, as well as
interaction with dogs and an unfamiliar person.
Acknowledgements
We thank all the dogs and humans who participated in the study and also Ulla Nilsson, Thomas
Gustavsson, and Sara Magnusson for assisting during the experiment. A special thank you to
Jan Hultgren for the statistical analysis.
313 Anthrozoös
Handlin et al.
AZ VOL. 24 (3).qxp:Layout 1 6/29/11 3:35 PM Page 313
References
Agren, G., Lundeberg, T., Uvnäs-Moberg, K. and Sato, A. 1995. The oxytocin antagonist 1-deamino-2-D-Tyr-
(Oet)-4-Thr-8-Orn-oxytocin reverses the increase in the withdrawal response latency to thermal, but not me-
chanical nociceptive stimuli following oxytocin administration or massage-like stroking in rats. Neuroscience
Letters 187(1): 49–52.
Allen, K., Blascovich, J. and Mendes, W. B. 2002. Cardiovascular reactivity and the presence of pets, friends,
and spouses: The truth about cats and dogs. Psychosomatic Medicine 64(5): 727–739.
Allen, K., Shykoff, B. E. and Izzo, J. L. Jr. 2001. Pet ownership, but not ace inhibitor therapy, blunts home blood
pressure responses to mental stress. Hypertension 38(4): 815–820.
Amico, J. A., Mantella, R. C., Vollmer, R. R. and Li, X. 2004. Anxiety and stress responses in female oxytocin
deficient mice. Journal of Neuroendocrinology 16(4): 319–324.
Araki, T., Ito, K., Kurosawa, M. and Sato, A. 1984. Responses of adrenal sympathetic nerve activity and cate-
cholamine secretion to cutaneous stimulation in anesthetized rats. Neuroscience 12(1): 289–299.
Arlt, W. and Stewart, P. M. 2005. Adrenal corticosteroid biosynthesis, metabolism, and action. Endocrinology
& Metabolism Clinics of North America 34(2): 293–313.
Carter, C. S. 1998. Neuroendocrine perspectives on social attachment and love. Psychoneuroendocrinology
23(8): 779–818.
Francis, D. D., Young, L. J., Meaney, M. J. and Insel, T. R. 2002. Naturally occurring differences in maternal care
are associated with the expression of oxytocin and vasopressin (V1a) receptors: Gender differences. Journal
of Neuroendocrinology 14(5): 349–353.
Friedmann, E. and Thomas, S. A. 1995. Pet ownership, social support, and one-year survival after acute
myocardial infarction in the Cardiac Arrhythmia Suppression Trial (CAST). American Journal of Cardiology
76(17): 1213–1217.
Handlin, L., Jonas, W., Petersson, M., Ejdeback, M., Ransjo-Arvidson, A. B., Nissen, E. and Uvnäs-Moberg, K.
2009. Effects of sucking and skin-to-skin contact on maternal ACTH and cortisol levels during the second
day postpartum-influence of epidural analgesia and oxytocin in the perinatal period. Breastfeeding Medicine
4(4): 207–220.
Hergovich, A., Monshi, B., Semmler, G. and Zieglmayer, V. 2002. The effects of the presence of a dog in the
classroom. Anthrozoös 15(1): 37–50.
Holst, S., Lund, I., Petersson, M. and Uvnäs-Moberg, K. 2005. Massage-like stroking influences plasma levels
of gastrointestinal hormones, including insulin, and increases weight gain in male rats. Autonomic
Neuroscience 120(1–2): 73–79.
Holst, S., Uvnäs-Moberg, K. and Petersson, M. 2002. Postnatal oxytocin treatment and postnatal stroking of
rats reduce blood pressure in adulthood. Autonomic Neuroscience 99(2): 85–90.
Jonas, W., Nissen, E., Ransjo-Arvidson, A. B., Wiklund, I., Henriksson, P. and Uvnäs-Moberg, K. 2008. Short- and
long-term decrease of blood pressure in women during breastfeeding. Breastfeeding Medicine 3(2): 103–109.
Kotrschal, K. and Ortbauer, B. 2003. Behavioral effects of the presence of a dog in the classroom. Anthrozoös
16: 147–159.
Kurosawa, M., Lundeberg, T., Agren, G., Lund, I. and Uvnäs-Moberg, K. 1995. Massage-like stroking of the
abdomen lowers blood pressure in anesthetized rats: Influence of oxytocin. Journal of the Autonomic Nervous
System 56(1–2): 26–30.
Kurosawa, M., Suzuki, A., Utsugi, K. and Araki, T. 1982. Response of adrenal efferent nerve activity to non-
noxious mechanical stimulation of the skin in rats. Neuroscience Letters 34(3): 295–300.
Liu, D., Diorio, J., Tannenbaum, B., Caldji, C., Francis, D., Freedman, A., Sharma, S., Pearson, D., Plotsky, P.
M. and Meaney, M. J. 1997. Maternal care, hippocampal glucocorticoid receptors, and hypothalamic-
pituitary-adrenal responses to stress. Science 277(5332): 1659–1662.
Lund, I., Ge, Y., Yu, L. C., Uvnäs-Moberg, K., Wang, J., Yu, C., Kurosawa, M., Agren, G., Rosen, A., Lekman,
M. and Lundeberg, T. 2002. Repeated massage-like stimulation induces long-term effects on nociception:
Contribution of oxytocinergic mechanisms. European Journal of Neuroscience 16(2): 330–338.
Lund, I., Lundeberg, T., Kurosawa, M. and Uvnäs-Moberg, K. 1999. Sensory stimulation (massage) reduces
blood pressure in unanaesthetized rats. Journal of the Autonomic Nervous System 78(1): 30–37.
Miller, S. C., Kennedy, C., DeVoe, D., Hickey, M., Nelson, T. and Kogan, L 2009. An examination of changes in
oxytocin levels in men and women before and after interaction with a bonded dog. Anthrozoös 22(1): 31–42.
Morelius, E., Theodorsson, E. and Nelson, N. 2005. Salivary cortisol and mood and pain profiles during skin-to-skin
care for an unselected group of mothers and infants in neonatal intensive care. Pediatrics 116(5): 1105–1113.
314 Anthrozoös
Short-Term Interaction between Dogs and Their Owners…
AZ VOL. 24 (3).qxp:Layout 1 6/29/11 3:35 PM Page 314
Nagasawa, M., Kikusui, T., Onaka, T. and Ohta, M. 2009. Dog’s gaze at its owner increases owner’s urinary
oxytocin during social interaction. Hormones and Behavior 55(3): 434–441.
Nissen, E., Lilja, G., Widstrom, A. M. and Uvnäs-Moberg, K. 1995. Elevation of oxytocin levels early post partum
in women. Acta Obstetricia et Gynecologica Scandinavica 74(7): 530–533.
Nissen, E., Uvnäs-Moberg, K., Svensson, K., Stock, S., Widstrom, A. M. and Winberg, J. 1996. Different
patterns of oxytocin, prolactin but not cortisol release during breastfeeding in women delivered by caesarean
section or by the vaginal route. Early Human Development 45(1–2): 103–118.
Odendaal, J. S. and Meintjes, R. A. 2003. Neurophysiological correlates of affiliative behaviour between humans
and dogs. The Veterinary Journal 165(3): 296–301.
Petersson, M., Alster, P., Lundeberg, T. and Uvnäs-Moberg, K. 1996. Oxytocin causes a long-term decrease
of blood pressure in female and male rats. Physiology and Behavior 60(5): 1311–1315.
Petersson, M., Hulting, A., Andersson, R. and Uvnäs-Moberg, K. 1999. Long-term changes in gastrin,
cholecystokinin and insulin in response to oxytocin treatment. Neuroendocrinology 69(3): 202–208.
Petersson, M., Hulting, A. L. and Uvnäs-Moberg, K. 1999. Oxytocin causes a sustained decrease in plasma
levels of corticosterone in rats. Neuroscience Letters 264(1–3): 41–44.
Richard, P., Moos, F. and Freund-Mercier, M. J. 1991. Central effects of oxytocin. Physiological Reviews 71(2):
331–370.
Serpell, J. 1991. Beneficial effects of pet ownership on some aspects of human health and behaviour. Journal
of the Royal Society of Medicine 84(12): 717–720.
Stock, S. and Uvnäs-Moberg, K. 1988. Increased plasma levels of oxytocin in response to afferent electrical
stimulation of the sciatic and vagal nerves and in response to touch and pinch in anaesthetized rats. Acta
Physiologica Scandinavica 132(1): 29–34.
Uvnäs-Moberg, K. 1996. Neuroendocrinology of the mother–child interaction. Trends in Endocrinology and
Metabolism: TEM 7(4): 126–131.
Uvnäs-Moberg, K. 1997. Physiological and endocrine effects of social contact. Annals of the New York Academy
of Sciences 807: 146–163.
Uvnäs-Moberg, K. 1998. Oxytocin may mediate the benefits of positive social interaction and emotions.
Psychoneuroendocrinology 23(8): 819–835.
Uvnäs-Moberg, K., Ahlenius, S., Hillegaart, V. and Alster, P. 1994. High doses of oxytocin cause sedation and low
doses cause an anxiolytic-like effect in male rats. Pharmacology, Biochemistry, and Behavior 49(1): 101–106.
Uvnäs-Moberg, K., Alster, P., Lund, I., Lundeberg, T., Kurosawa, M. and Ahlenius, S. 1996. Stroking of the
abdomen causes decreased locomotor activity in conscious male rats. Physiology and Behavior 60(6):
1409–1411.
Uvnäs-Moberg, K. and Eriksson, M. 1996. Breastfeeding: Physiological, endocrine and behavioural adapta-
tions caused by oxytocin and local neurogenic activity in the nipple and mammary gland. Acta Paediatrica
85(5): 525–530.
Uvnäs-Moberg, K., Handlin, L. and Petersson, M. 2011. Promises and pitfalls of hormone research in human–
animal interaction. In How Animals Affects Us: Examining the Influence of Human–Animal Interaction on Child
Development and Human Health, 53–81, ed. P. MacCardle, S. McCune, J. A. Griffin and V. Maholmes. Wash-
ington, DC: American Psychological Association.
Uvnäs-Moberg, K., Lundeberg, T., Bruzelius, G. and Alster, P. 1992. Vagally mediated release of gastrin and
cholecystokinin following sensory stimulation. Acta Physiologica Scandinavica 146(3): 349–356.
Uvnäs-Moberg, K. and Petersson, M. in press. Role of oxytocin and oxytocin related effects in manual therapies.
In The Science and Clinical Application of Manual Therapy, ed. H. King, W. Jänig and M. M. Patterson.
Elsevier: Amsterdam.
Uvnäs-Moberg, K., Widström, A. M., Nissen, E. and Björvell, H. 1990. Personality traits in women 4 days
postpartum and their correlation with plasma levels of oxytocin and prolactin. Journal of Psychosomatic
Obstetrics and Gynecology (11): 261–273.
Wilson, C. C. 1991. The pet as an anxiolytic intervention. The Journal of Nervous and Mental Disease 179(8):
482–489.
315 Anthrozoös
Handlin et al.
AZ VOL. 24 (3).qxp:Layout 1 6/29/11 3:35 PM Page 315
