Content uploaded by Benjamin L Hart
Author content
All content in this area was uploaded by Benjamin L Hart on May 02, 2015
Content may be subject to copyright.
T
hanks to improvements in nutrition, medical care,
and protection from accidental death, many com-
panion dogs live well into their elderly stages of life. As
of 1996, 14% (7.3 million) of all dogs in the United
States were 11 years of age or older.
1
A longer life span
brings with it age-related degenerative changes in a
number of organ systems, including those of the car-
diovascular, musculoskeletal, visual, auditory, renal,
and endocrine systems. The brain also undergoes
degenerative changes with age, resulting in noticeable
impairment of memory and learning in dogs.
2,3
The
growing interest in understanding the nature and
prevalence of signs of dysfunction in dogs related to
impairment of memory and learning was enhanced by
the recent approval of the dopaminergic and neuropro-
tective drug selegiline hydrochloride
a
by the FDA
Center for Veterinary Medicine for treatment of cogni-
tive dysfunction syndrome in dogs.
4
Because signs such as disorientation in the home,
disturbance of house training, and reduced recognition
of family members involve deficits in memory and
learning, the terms cognitive dysfunction and cognitive
impairment have been introduced into the literature
dealing with aging dogs to refer to these behavioral
changes.
2,3,5
Animal cognition usually refers to complex
learned associations where the animal makes respons-
es that appear to reflect abstractions from previously
learned material; that is, a new response is made that
was not the result of prior direct reinforcement.
6,7
Although such cognitive function would be difficult to
assess clinically in companion dogs, this species, like
many mammalian and avian species studied,
6,7
is
undoubtedly capable of cognitive behavior. Loss of ori-
entation in the home or yard, impaired recognition of
family members, and loss of house training, although
not necessarily manifestations of impairment in cogni-
tive behavior, do reflect impairment of memory and
learning and would be logically linked with impair-
ment of cognitive function.
Use of the terms cognitive dysfunction or cogni-
tive impairment is predicated on the assumption that
these changes reflect neuronal, neurotransmitter, meta-
bolic, cerebrovascular, or biochemical degenerative
changes in the brain and are not primarily a reflection
of dysfunction in other organ systems.
2,3
Aged dogs
have been known for some time to accumulate β-amy-
loid deposits in the hippocampus and frontal cortex
(areas especially involved in changes in cognitive
behavior),
8,9
and results of recent research indicate that
the deposits are similar to the primitive or early stage
β-amyloid plaques in the brains of human Alzheimer’s
patients.
10,11
In humans with Alzheimer’s disease, the
plaques mature into discrete plaques that result in neu-
ronal cell death.
12,13
Results of laboratory research on
aging Beagles suggest that the degree of β-amyloid
JAVMA, Vol 218, No. 11, June 1, 2001 Scientific Reports: Original Study 1787
SMALL ANIMALS
Prevalence of behavioral changes associated
with age-related cognitive impairment in dogs
Jacqueline C. Neilson, DVM, DACVB; Benjamin L. Hart, DVM, PhD, DACVB; Kelly D. Cliff, DVM;
William W. Ruehl,
DVM, PhD, DACVP
Objective—To determine the prevalence of age-relat-
ed behavioral changes, namely impairment, in a ran-
domly chosen population of dogs.
Design—Age-stratified cohort study.
Animals—97 spayed female and 83 castrated male
dogs that were 11 to 16 years old.
Procedure—Data on possible impairment in 4 behav-
ioral categories (ie, orientation in the home and yard,
social interaction, house training, and sleep-wake cycle)
linked to cognitive dysfunction were obtained from dog
owners, using a structured telephone interview.
Hospital records of dogs had been screened to exclude
dogs with dysfunction in organ systems that may cause
behavioral changes. Dogs with behavioral impairment
were those with ≥ 2 signs of dysfunction within a cate-
gory. Dogs with impairment in 1 category were consid-
ered mildly impaired and those with impairment in ≥ 2
categories were considered severely impaired.
Results—Age by sex interactions for dogs with
impairment in any category were not significant, and,
therefore, data on castrated males and spayed
females were pooled for analyses across ages. The
prevalence of age-related progressive impairment
was significant in all categories. The percentage of
11- to 12-year-old dogs with impairment in ≥ 1 cate-
gory was 28% (22/80), of which 10% (8/80) had
impairment in ≥ 2 behavioral categories. Of 15- to 16-
year-old dogs, 68% (23/34) had impairment in ≥ 1 cat-
egory, of which 35% (12/34) had impairments in ≥ 2
categories. There were no significant effects of body
weight on the prevalence of signs of dysfunction in
the behavioral categories.
Conclusions and Clinical Relevance—Data collect-
ed provide estimates of the prevalence of various
degrees of age-related behavioral changes associated
with cognitive dysfunction in dogs. Age-related behav-
ioral changes may be useful indicators for medical
intervention for dogs with signs of cognitive impair-
ment. (
J Am Vet Med Assoc
2001;218:1787–1791)
From the Animal Behavior Clinic, 809 SE Powell, Portland, OR
97202 (Neilson); and the Behavior Service, Veterinary Medical
Teaching Hospital (Hart), and the Department of Anatomy,
Physiology, and Cell Biology (Cliff), School of Veterinary
Medicine, University of California, Davis, CA 95616; and 13868
Skyline Blvd, Woodside, CA 94062 (Ruehl).
Supported in part by Friskies PetCare Residency Program in Clinical
Animal Behavior at the University of California, Davis, the UC
Davis Center for Companion Animal Health, and Deprenyl Animal
Health, Overland Park, Kan.
The authors thank Dr. David Bruyette for technical assistance and
Dr. Neil Willits for statistical assistance.
Address correspondence to Dr. Hart.
deposition corresponds to degree of impairment in
learning complex tasks.
14,15
Neuropathologic and
behavioral changes described in some aging dogs are
similar to behavioral changes in humans with demen-
tia and suggest that canine dementia can serve as a
model for human dementia.
5
A further link between age-related behavioral
changes in dogs and of humans with signs of
Alzheimer’s disease was profiled in 2 double-blind con-
trolled studies with the drug selegiline, 1 revealing a
slowing in the progression of degenerative changes in
human Alzheimer’s patients,
16
and the other revealing a
significant improvement in aged dogs with cognitive
dysfunction syndrome.
4
Although dogs can be expected to have some aging
of behavioral function, it could be that only a propor-
tion of dogs at any age will have cerebral degeneration
of the magnitude sufficient to cause severe impair-
ments in memory and learning. Owners of older dogs
may describe a variety of behavioral changes, and
results of preliminary work indicated that aside from a
reduction in activity, the signs could generally be
grouped into the following 4 categories
2,3
: orientation
in the immediate environment, social interactions with
human family members, house training, and the sleep-
wake cycle. Changes in the sleep-wake cycle are per-
haps the most difficult to relate to deficits in memory
or learning. However, the sleep-wake cycle is disturbed
in humans with Alzheimer’s, possibly as a reflection of
disruption of the circadian rhythm.
13,17
The purpose of the study presented here was to
assess the age-related prevalence of behavioral changes
in a randomly selected population of aging dogs receiv-
ing veterinary care. In addition to the behavioral
changes, information was collected on visual and hear-
ing impairment. It is known that dogs of larger body size
physically age more rapidly than dogs of smaller body
size.
18
Thus, we also set out to determine whether there
was a correspondence between body weight and the age
of onset of signs of failing learning and memory.
Materials and Methods
Selection of dogs—Dogs included in our study were
receiving veterinary care and were chosen from the database
of clients of the Veterinary Medical Teaching Hospital at the
University of California, Davis. A computer program was
used to randomly generate lists of case numbers of dogs that
fell into targeted age groups.
Data collection—The hospital records were reviewed by
a trained screener. Clients were then contacted by the screen-
er to obtain information about whether their dogs were still
alive and the willingness of the caregiver to participate in the
study. Information was also sought about the existence of
medical problems not apparent from the hospital record and a
convenient time for a veterinary behaviorist (JCN) to call. The
owner was then sent an information sheet about possible
behavioral and physical changes in aging dogs, and approxi-
mately 2 weeks later, the behaviorist contacted the owner to
conduct a structured telephone interview of approximately 30
minutes. An outside consultant on telephone interview tech-
niques contributed to development of the interview methods
to avoid biasing the responses of the interviewee.
Information was recorded on prepared data forms and
organized onto a spreadsheet. Data collection forms dealt
with the following 4 categories of behaviors: orientation in
the home or yard, social interactions with human family
members, house training, and the sleep-wake cycle. Prior to
asking about signs of dysfunction in each category, there was
an open-ended question for recording responses from each
owner about signs displayed by their dogs. Following the
open-ended question, the interviewer then asked about spe-
cific changes known to develop in older dogs associated with
each category if these signs had not been previously men-
tioned by the owner. The initial part of the interview con-
tained distractors to intentionally evoke comments of “no
change” from the owner so they would not expect that ques-
tions should be answered in the affirmative.
During the interview, 3 to 4 specific questions were used
that addressed signs of dysfunction in each behavioral cate-
gory. The criterion for impairment in a category was that
dogs had ≥ 2 distinct signs in that category that were not
observed when they were younger (ie, 5 to 8 years of age). A
dog had impairment in orientation if it had ≥ 2 of the fol-
lowing signs: staring into space, getting lost in the house or
yard, getting stuck in corners, standing at the wrong door or
wrong part of the door to go out, and any other sign that was
logically attributable to disorientation. A dog had impair-
ment in social interaction if it had ≥ 2 of the following signs:
decrease in greeting owners, decrease in soliciting attention
from the owners, and a change (increase or decrease) in fol-
lowing owners around the house. A dog had impairment in
house training if it had the following 2 signs: started to uri-
nate or defecate in the house with no other behavioral or
medical explanation discernible such as urinary incontinence
or separation anxiety, and a decrease in signaling to go out or
a decrease in use of a dog door. If a dog was strictly an out-
door dog or never house trained, so that learned house train-
ing could not be evaluated, the dog was not included in our
study. A dog had impairment in the sleep-wake cycle if it had
≥ 2 of the following signs: regularly waking up the owner at
night by pacing or vocalizing, sleeping less at night, and
sleeping noticeably more during the day.
Evaluation system—To have impairment in a category, a
dog had to have the signs of dysfunction in that category ≥ 1
time a week continuously for at least the previous month. The
system was considered conservative. For example, it may be rea-
sonable to assume that a dog had disorientation if it regularly
stared into space, but a dog was not considered to have impair-
ment in orientation unless there was ≥ 1 other marker of disori-
entation. Owners were asked about visual impairment and hear-
ing impairment that was not noticeable earlier in the dog’s life,
and this information was considered during evaluation.
Owners were also asked about behavioral problems such
as recent onset of separation anxiety, fears of inanimate stim-
uli, and aggression. If these problems could account for
behavioral findings, the dog was not considered to have
impairment in the related category.
The primary analysis dealt with each of the 4 behavioral
categories. It was determined whether each dog had impair-
ment in 0, ≥ 1, and ≥ 2 categories. Dogs with impairment in
1 category were arbitrarily considered to have mild impair-
ment, whereas dogs with impairment in ≥ 2 categories were
arbitrarily considered to have severe impairment.
Other methods of designating severity of impairment
were considered, such as giving more weight to impairment
in some categories than others (eg, requiring disorientation
for a classification of severe impairment). Analysis of pilot
data, however, revealed that the prevalence of dogs in each
age group with impairments was not affected by whether data
were weighted or not.
Statistical analysis—Analyses of changes in the propor-
tion of dogs with impairments in the various categories as well
1788 Scientific Reports: Original Study JAVMA, Vol 218, No. 11, June 1, 2001
SMALL ANIMALS
as visual and hearing impairment across age groups included
not only the possibility of sex differences but also the possibil-
ity of interaction between sex and age, which would compli-
cate an analysis of the progression of signs of dysfunction
across the age groups. The significance of such interactions
was examined by logistic regression. Examination of age
effects, after suppressing sex as a factor, was done with the
Fisher exact test. The analysis of the degree of impairment as
a function of body weight involved the possibility of interac-
tion between body weight and age, which was also addressed
by logistic regression. The effect of weight alone was examined
after ruling out the age by weight interaction. Significance for
all analyses was set at P < 0.05, using a 2-tailed t-test.
Results
Dogs—Data were obtained on 180 dogs, compris-
ing 83 castrated males and 97 spayed females and rep-
resenting 3 age groups, namely, 11 to 12 (n = 80), 13 to
14 (66), and 15 to 16 (34) years old. These were dogs
for which there was no evidence in the hospital record
that could provide an explanation for behavioral
changes that could be attributed to dysfunction in ≥ 1
organ system other than the CNS.
Data were available for a somewhat larger number of
dogs in the younger age groups than in the oldest group.
Of the 11- to 12-year-old dogs, 56% (45/80) were females
that were spayed at a mean age of 2.9 (± 2.6) years, and
44% (35/80) were males that were castrated at a mean
age of 3.6 (± 3.1) years. Of the 13- to 14-year-old dogs,
53% (35/66) were females that were spayed at a mean age
of 2.7 (± 2.6) years, and 47% (31/66) were males that
were castrated at a mean age of 5.3 (± 4.1) years. Of the
15- to 16-year-old dogs, 50% (17/34) were females that
were spayed at a mean age of 2.3 (± 2.6) years, and 50%
(17/34) were males that were castrated at a mean age of
6.3 (± 4.2) years. Therefore, the percentage of each sex in
each of 3 age groups was approximately the same.
Overall, females were spayed at a mean age of 2.7 (± 2.6)
years, and males were castrated at a mean age of 4.8
(± 3.8) years. Mean elapsed time from the last hospital
visit indicated on the hospital record and the data-collec-
tion telephone call was 20.6 (± 22.5) months.
There was not a significant effect of age by sex
interaction on the 4 categories and the 3 degrees of
impairment represented by impairment in ≥ 1 or ≥ 2
categories (P values ranged from 0.26 to 0.85). To
determine possible sex differences in proportion of
dogs with impairments in the various categories and
degrees of impairment, age groups were combined (Fig
1). Castrated male dogs were significantly more likely
have impairment in orientation than spayed female
dogs (P = 0.03). In the other 3 categories and 3 degrees
of impairment, there were no significant sex differ-
ences (P values ranged from 0.06 to 0.35).
Age-related prevalence of impairment in behav-
ioral categories—Because there were no significant sex
by age interactions, subsequent analyses used com-
bined data from female and male dogs across age groups
(Fig 2). There was a progressive and significant age-
related prevalence of dogs with impairment in each cat-
egory and proportion of dogs with impairments in ≥ 1
and ≥ 2 categories (P values ranged from 0.001 to
0.004). The proportion of dogs with impairment in just
1 category was not expected to be significantly different
between age groups, because further impairment with
advancing age would be expressed as impairment in ≥ 2
categories. The percentage of 11- to 12-year-old dogs
with impairment in ≥ 1 category was 28% (22/80), of
which 10% (8/80) had impairment in ≥ 2 categories.
Among 15- to 16-year-old dogs, 68% (23/34) had
impairment in ≥ 1 category, of which 35% (12/34) had
impairments in ≥ 2 categories. There was no indication
of a chronologic progression of impairment such that
dogs would have impairment in any particular category
prior to developing impairment in other categories.
Impairment of vision and hearing—Visual and
hearing impairment was described by the owner and
not necessarily evaluated by a veterinarian. There were
no sex by age interactions for visual or hearing impair-
ment (P = 0.39 and 0.45, respectively). Combining
data from males and females for further analysis
revealed a significant age-related increase in prevalence
of visual and hearing impairment (P = 0.03 and P <
0.001, respectively). Visual impairment progressed
JAVMA, Vol 218, No. 11, June 1, 2001 Scientific Reports: Original Study 1789
SMALL ANIMALS
Figure 1—Percentage of male castrated (MC; n = 83) and
female spayed (FS; 97) dogs that had impairments in each of 4
behavioral categories and in 1 category, ≥ 2 categories, and ≥ 1
category.
a,b
Different letters within a set of bar graphs (ie, 2 bars)
indicate a significant difference (
P
< 0.05) between sexes.
Figure 2—Percentage of 11- to 12-year-old (n = 80), 13- to 14-
year-old (66), and 15- to 16-year-old (34) dogs that had impair-
ments in each of 4 behavioral categories and in 1 category, ≥ 2
categories, and ≥ 1 category.
a,b
Different letters within a set of
bar graphs indicate significant differences (
P
< 0.05) among age
groups.
from 41% (33/80) of 11- to 12-year-old dogs to 68%
(23/34) of 15- to 16-year-old dogs. Hearing impair-
ment progressed from 48% (38/80) of 11- to 12-year-
old dogs to > 97% (33/34) of 15- to 16-year-old dogs.
Body weight and signs of dysfunction—The issue
of body weight and predisposition for behavioral
impairment was addressed by testing for interactions
between age and weight for dogs with impairment in
≥ 2 categories. The interaction was not significant, and
the analysis conducted without the age by weight inter-
action revealed no significant weight effect (P = 0.36).
Discussion
Our cross-sectional analysis of data from dogs
ranging in age from 11 to 16 years provides information
on the prevalence of age-related behavioral changes
associated with cognitive dysfunction. These signs were
the result of impairment in the behavioral categories of
orientation in the house and yard, social interactions
with human family members, house training, and the
sleep-wake cycle. Impairment in 1 category was arbi-
trarily designated as mild impairment, and impairment
in ≥ 2 categories was designated as severe impairment.
As discussed, signs of dysfunction in all the behavioral
categories except the sleep-wake cycle can be related, at
least in part, to disturbances of learning and memory.
Although the behavioral patterns affected may not rep-
resent cognitive function as studied by animal behav-
iorists or neuroscientists, the changes in dogs corre-
spond with neuropathologic changes, especially the
deposition of diffuse β-amyloid plaques,
14,15
and share
similarities with Alzheimer’s disease in humans, which
is also characterized by β-amyloid deposition.
12,13,17
Behavioral impairments in some dogs in our study
may have actually been a reflection of an organ system
abnormality (other than brain degeneration). More
likely, perhaps, is the possibility that hypothyroidism,
arthritis, and cardiovascular disease may have intensi-
fied some signs that also reflected loss of memory and
learning. In fact, most old dogs have ≥ 1 major organ
system disorder, so it would be unusual to find a 15- to
16-year-old dog with impairment in ≥ 2 behavioral cat-
egories without any substantial impairment in any
nonbrain organ system. Data from the clinical trial
with selegiline,
4
involving 199 dogs that had an exten-
sive physical examination with a routine neurologic
examination, hematologic analysis, blood biochemical
analysis (including T
4
analysis), and urinalysis to rule
out medical disorders as causal factors in the signs of
cognitive dysfunction, provided evidence that the signs
of dysfunction in the categories of orientation, social
interaction, house training, and sleep-wake cycle were
not typically a consequence of a medical disorder
(other than brain degeneration). We believe that our
demographic study is a fair approximation of the age-
related prevalence of the signs associated with cogni-
tive dysfunction in dogs.
Further confirmation of the legitimacy of our
demographic study as a guide to age-related prevalence
of this syndrome comes from the longitudinal study
19
from this institution focusing on the short-term pro-
gression of impairment in a subset of dogs of our study
presented here. Results of that study indicate that
almost all dogs that initially had impairment in a cate-
gory continued to have impairment in the category,
and many progressed to a more severe degree of
impairment, because they developed impairment in ≥ 1
additional category during a subsequent 6- to 18-
month period. If the signs used to define impairments
in our study presented here had reflected disturbances
in the environment, interactions with the owner, or
transient medical problems, rather than age-related
pathologic changes in the brain, one would expect a
reversal in many of the signs, and this would have
invalidated the particular behavioral measures used.
Results from the longitudinal study
19
also indicate
that the onset of visual impairment, hearing impair-
ment, or both accounted for only a small proportion of
dogs that developed disorientation during the same
period. Thus, although some disorientation developed
or was intensified by visual or hearing impairment,
most instances of disorientation were independent of
impairment in these organ systems.
In this study, 23% (22/80 of 11- to 12-year olds
and 68% (23/34) of 15- to 16-year-olds had impair-
ment in ≥ 1 category. Ten percent (8/80) of 11- to 12-
year-olds and 35% (12/34) of 15- to 16-year-olds had
impairment that was designated as severe in ≥ 2 cate-
gories. The only significant difference between male
and female dogs was a higher proportion of castrated
male than spayed female dogs with impairment in ori-
entation. Visual and hearing impairment also increased
with age; 68% (23/34) of 15- to 16-year-old dogs had
visual impairment, and 97% (33/34) had hearing
impairment. The age-related increase in impairment in
dogs for all behavioral categories was significant, as
was the increase in impairment in visual and hearing
function with advancing age.
Although it is established that dogs of large body
size die at a younger age, presumably as a function of
more rapid aging in the cardiovascular, musculoskele-
tal, and endocrine systems,
18
findings of our study sug-
gest that the processes in the brain responsible for age-
related behavioral changes do not progress more rapid-
ly in larger dogs. At any given age, small dogs are as
likely to have impairment in ≥ 2 categories as large
dogs. This means that because small dogs often live
longer, one would predict they would be more likely to
have signs associated with cognitive dysfunction dur-
ing their lifetime than dogs of larger body size.
If impairment in ≥ 1 category is an indication for
medical intervention, results of our study indicate that
almost 30% of dogs 11 to 12 years of age would be can-
didates for such treatment, as would about 70% of dogs
15 to 16 years of age. On the assumption that dogs often
progress in severity of signs of impairment in behavioral
categories with advancing age, some clinicians may wish
to recommend treatment when a dog has 1 sign of dys-
function in a behavioral category, as long as an extensive
medical evaluation precludes an organ system disorder
as the causal factor (other than brain degeneration). It
should also be emphasized that there are undoubtedly
other behavioral changes associated with aging that
were not included in our study, which could be used as
indications for medical interventions.
1790 Scientific Reports: Original Study JAVMA, Vol 218, No. 11, June 1, 2001
SMALL ANIMALS
a
Anipryl, Pfizer Animal Health, Exton, Pa.
References
1. American Veterinary Medical Association. US pet ownership
and demographics sourcebook. Schaumburg, Ill: Center for
Information Management, American Veterinary Medical Association,
1997.
2. Ruehl WW, Hart BL. Canine cognitive dysfunction. In:
Dodman N, Shuster L, eds. Psychopharmacology of animal behavior dis-
orders. Malden, Mass: Blackwell Scientific Publications, 1998;283–304.
3. Ruehl WW, Bruyette DS, De Paoli A, et al. Canine cognitive
dysfunction as a model for human age-related cognitive decline,
dementia and Alzheimer’s disease: clinical presentation, cognitive
testing, pathology and response to l-deprenyl therapy. Prog Brain Res
1995;106:217–225.
4. FDA. Freedom of information summary, Anipryl (selegiline
hydrochloride), Pfizer, Inc, Groton, CT. Washington, DC: US FDA, 1998.
5. Cummings BJ, Head E, Ruehl WW, et al. The canine as an
animal model of human aging and dementia. Neurobiol Aging
1996;17:259–268.
6. Shettleworth SJ. Cognition, evolution and behavior. New
York: Oxford University Press, 1998;3–48.
7. Czeschi T. Animal cognition—the phylogeny of cognitive
abilities. Anim Cogn 1998;1:1–2.
8. Wisniewski H, Johnson AB, Raine CS, et al. Senile plaques
and cerebral amyloidosis in aged dogs. A histochemical and ultra-
structural study. Lab Invest 1970;23:287–296.
9. Uchida K, Miyauchi Y, Kakayama H, et al. Amyloid
angiopathy with cerebral hemorrhage and senile plaques in aged
dogs. Jpn J Vet Sci 1990;52:605–611.
10. Cummings B, Su JH, Cotman CW, et al. β-Amyloid accu-
mulation in aged canine brain. A model of early plaque formation in
Alzheimer’s disease. Neurobiol Aging 1993;14:547–560.
11. Cummings BJ, Satou T, Head E, et al. Diffuse plaques con-
tain C-terminal Aβ
42
and not Aβ
40
: evidence from cats and dogs.
Neurobiol Aging 1996;17:653–659.
12. Cummings BJ, Cotman DW. Image analysis of β-amyloid
load in Alzheimer’s disease and relation to dementia severity. Lancet
1995;346:1524–1528.
13. Ashford JW, Schmitt F, Kumar V. Diagnosis of Alzheimer’s
disease. In: Kumar V, Eisdorfer C, eds. Advances in the diagnosis and
treatment of Alzheimer’s disease. New York: Springer Publishing Co,
1998;111–136.
14. Cummings BJ, Head E, Afagh AJ, et al. β-Amyloid accumu-
lation correlates with cognitive dysfunction in the aged canine.
Neurobiol Learn Mem 1996;66:11–23.
15. Head E, Callahan H, Muggenburg BA, et al. Visual-discrim-
ination learning ability and β-amyloid accumulation in the dog.
Neurobiol Aging 1998;19:415–425.
16. Sano M, Ernesto C, Thomas RG, et al. A controlled trial of
selegiline, alpha-tocopherol, or both as treatment for Alzheimer’s dis-
ease. N Engl J Med 1997;336:1216–1222.
17. American Psychiatric Association. Diagnostic and statistical
manual of mental disorders. 4th ed. Washington, DC: American
Psychiatric Association, 1994:133–155.
18. Patronek GJ, Waters DJ, Glickman LT. Comparison longevi-
ty of pet dogs and humans: implications for gerontology research. J
Gerontol A Biol Sci Med Sci 1997;52:B171–B178.
19. Bain MJ, Hart BL, Cliff, KD, et al. Predicting behavioral
changes associated with age-related cognitive impairment in dogs. J
Am Vet Med Assoc 2001;218:1792–1795.
JAVMA, Vol 218, No. 11, June 1, 2001 Scientific Reports: Original Study 1791
SMALL ANIMALS
