1102 Scientific Reports: Original Study JAVMA, Vol 227, No. 7, October 1, 2005
accines are developed to be immunogens and are
required to have potency, safety, and efficacy before
licensing; however, no vaccine is absolutely reaction free
or completely effective. Although manufacturers’ pre-
marketing safety trials help ensure that vaccine-associ-
ated adverse events (VAAEs) occur infrequently, VAAEs
have evoked public and professional concern regarding
the possible overvaccination of humans and animals.
Traditional postmarketing surveillance of veteri-
nary vaccines relies on veterinarians and owners to
voluntarily submit reports of suspected reactions to
manufacturers or the USDA. Limitations to passive sur-
veillance include variability in report quality, selective
reporting, underreporting, and the inability to deter-
mine whether a vaccine caused the adverse event
described in any report.
The information submitted
in voluntarily generated reports is not necessarily com-
plete and often lacks important details about the ani-
mal, observed event, or concurrent medications
received by the affected animal. Incidence rates and
relative risks for specific VAAEs cannot be calculated
because of the lack of information about the overall
vaccinated population (denominator data or the popu-
lation at risk of a VAAE).
Large, population-based medical record databases
have been used in recent years to conduct epidemio-
logic investigations of human vaccine safety.
these databases are usually generated during the rou-
tine administration of medical care and do not require
the completion of a separate VAAE form, the problem
of underreporting of events is reduced. Large popula-
tions can be examined for infrequent adverse events,
and denominator data on doses given and the avail-
ability of appropriate comparison groups make these
databases ideal for studying vaccine safety.
Banfield, The Pet Hospital, is a small animal gen-
eral practice that uses the same computerized medical
record system in more than 400 hospital locations
throughout the United States. Electronic medical
records are stored in a central data warehouse and can
be easily retrieved for administrative or medical review
and analysis. Proprietary software
dardized codes for physical examinations, laboratory
tests, diagnoses, and treatments is used by all veteri-
narians in the practice. The codes, when combined
with dates, allow determination of temporal relation-
ships between vaccination and potential VAAEs, a fac-
tor that is important for assessing causality.
Abnormalities noticed within a few days of vaccination
may be attributable to an adverse response to the vac-
cine immunogens, but the risk factors and incidence of
these events in companion animals have not been char-
acterized. The purpose of the study reported here was
to use the Banfield database to estimate the incidence
rate and potential risk factors for VAAEs that occurred
within 3 days of vaccine administration in dogs.
Adverse events diagnosed within three days
of vaccine administration in dogs
George E. Moore, DVM, MS, PhD, DACVPM, DACVIM; Lynn F. Guptill, DVM, PhD, DACVIM;
Michael P. Ward,
BVSc, MS, MPVM, PhD; Nita W. Glickman, MPH, PhD; Karen K. Faunt, DVM, DACVIM;
Hugh B. Lewis,
BVMS, DACVP; Lawrence T. Glickman, VMD, DrPH
Objective—To determine incidence rates and poten-
tial risk factors for vaccine-associated adverse events
(VAAEs) diagnosed within 3 days of administration in
Design—Retrospective cohort study.
Animals—1,226,159 dogs vaccinated at 360 veteri-
Procedure—Electronic records from January 1, 2002,
through December 31, 2003, were searched for pos-
sible VAAEs (nonspecific vaccine reaction, allergic
reaction, urticaria, or anaphylaxis) diagnosed within 3
days of vaccine administration. Information included
age, weight, sex, neuter status, and breed. Specific
clinical signs and treatments were reviewed in a ran-
dom sample of 400 affected dogs. The association
between potential risk factors and a VAAE was esti-
mated by use of multivariate logistic regression.
Results—4,678 adverse events (38.2/10,000 dogs
vaccinated) were associated with administration of
3,439,576 doses of vaccine to 1,226,159 dogs. The
VAAE rate decreased significantly as body weight
increased. Risk was 27% to 38% greater for neutered
versus sexually intact dogs and 35% to 64% greater
for dogs approximately 1 to 3 years old versus 2 to 9
months old. The risk of a VAAE significantly increased
as the number of vaccine doses administered per
office visit increased; each additional vaccine signifi-
cantly increased risk of an adverse event by 27% in
dogs ≤ 10 kg (22 lb) and 12% in dogs > 10 kg.
Conclusions and Clinical Relevance—Young adult
small-breed neutered dogs that received multiple vac-
cines per office visit were at greatest risk of a VAAE
within 72 hours after vaccination. These factors
should be considered in risk assessment and risk
communication with clients regarding vaccination.
J Am Vet Med Assoc
From the Departments of Veterinary Pathobiology (Moore, Ward,
Glickman, Glickman) and Veterinary Clinical Sciences (Guptill),
School of Veterinary Medicine, Purdue University, West Lafayette,
IN 47907-2027; and Banfield, The Pet Hospital, 11815 NE Glenn
Widing Dr, Portland, OR 97220 (Faunt, Lewis). Dr. Ward’s present
address is the Department of Veterinary Integrative Biosciences,
College of Veterinary Medicine and Biomedical Sciences, Texas
A&M University, College Station, TX 77843-4458.
Supported by the CDC National Center for Infectious Diseases
(grant No. R01 CI000093).
Address correspondence to Dr. Moore.
05-02-0089.qxp 9/13/2005 10:38 AM Page 1102
JAVMA, Vol 227, No. 7, October 1, 2005 Scientific Reports: Original Study 1103
Materials and Methods
Electronic medical records of all dogs at 360 Banfield vet-
erinary hospitals from January 1, 2002 through December 31,
2003 were extracted from the Banfield practice database.
Records were included in the study if the species code was
canine and if treatments included a code for Bordetella vac-
parainfluenza-parvovirus-leptospirosis (serovars canicola,
icterohaemorrhagiae, grippotyphosa, and pomona) vaccine,
Records for dogs that received both an
injectable heartworm preventive and a vaccine during the
same office visit were not included in analyses.
The date of birth, breed, sex, neuter status, weight, and
date of vaccination were recorded for each dog. Medical
records of dogs that received vaccinations were searched for
possible VAAEs by use of diagnostic codes (vaccine reaction,
allergic reaction, urticaria, anaphylaxis, cardiac arrest, cardio-
vascular shock, and sudden death). Diagnoses
were only included in analyses if the diagnosis
date was within 3 days after vaccine administra-
tion. Date of death, if recorded, was used to
determine if death occurred within 3 days of vac-
cination. The free-text medical note field was
reviewed in a random sample of 400 affected
dogs to determine clinical signs and treatments
associated with a VAAE.
Statistical analyses—All calculations were
performed with statistical software.
neuter status were analyzed as categoric vari-
ables. Continuous variables of age and weight
were converted to categoric variables because
nonlinear trends were detected in the model-fit-
ting process. Dogs were grouped on the basis of
age at date of vaccination as follows: 2 to 9
months, > 9 months to 1.5 years, > 1.5 to 2.5
years, > 2.5 to 3.5 years, > 3.5 to 5.5 years, > 5.5
to 8.5 years, and > 8.5 years. Weight was con-
verted from a continuous to a categoric variable
of 5-kg (11-lb) weight groups up to 45 kg
(99 lb). Incidence rates with 95% confidence
limits were calculated via assumption of a bino-
mial distribution for proportions. Tests for trend
were performed across ordered groups by use of
the Cuzick nonparametric test. All VAAE rates
are reported as the number of adverse
events/10,000 dogs vaccinated. For categoric
variables, rates for affected dogs were compared
with rates for nonaffected dogs by use of the χ
test for independence.
Potential risk factors for VAAEs were
evaluated by use of bivariate and multivariate
unconditional logistic regression, and multi-
variate logistic regression included dog ran-
dom effects. Estimates of the odds ratio (OR)
and 95% confidence intervals (CIs) for each
risk factor were obtained by use of exponen-
tials of the regression coefficients. Interactions
between independent variables in the final
multivariate model were assessed for an asso-
ciation with an adverse event. Interactions
between vaccines were not included in multi-
variate analysis because of the large number of
vaccine combinations (n = 58). Maximum
likelihood estimates of the logistic parameters
and final model were assessed for significance
by use of the Hosmer-Lemeshow χ
of-fit test. A value of P < 0.05 was considered
Population—In the 2-year study period, 4,531,837
vaccine doses were administered to 1,537,534 dogs at
360 veterinary hospitals. Excluding from analysis
311,375 dogs that concurrently received injectable
heartworm preventive and vaccinations, 3,439,576 vac-
cine doses were administered to 1,226,159 dogs. Mean
number of vaccine doses administered per dog per
office visit was 2.8 (range, 1 to 6), and 4,678 dogs had
a diagnosis of a VAAE within 3 days of vaccine admin-
istration (38.2 VAAEs/10,000 dogs vaccinated; 95% CI,
37.1 to 39.3). The percentage of VAAEs diagnosed on
days 0 (same day), 1, 2, and 3 postvaccination was
72.8%, 18.9%, 5.5%, and 2.8%, respectively. Of 4,678
VAAEs, 3,080 (65.8%) were coded as vaccine reactions,
1,481 (31.7%) as allergic reactions, 80 (1.7%) as ana-
Figure 1—Mean ± SEM vaccine-associated adverse event (VAAE) rates (No. of
adverse events/10,000 dogs vaccinated) by 5-kg (11-lb) weight groups in 1,226,159
dogs vaccinated at 360 veterinary hospitals from January 1, 2002, to December 31,
2003. The VAAEs were diagnosed within 3 days of vaccine administration.To convert
kilograms to pounds, multiply by 2.2.
Figure 2—Mean ± SEM VAAE rates for all vaccines (n = 1,226,159) and for rabies
vaccine administered alone (118,765) to dogs of various body weights (BWs).
Figure 1 for remainder of key.
05-02-0089.qxp 9/13/2005 10:38 AM Page 1103
1104 Scientific Reports: Original Study JAVMA, Vol 227, No. 7, October 1, 2005
Figure 3—Mean ± SEM VAAE rates by age in dogs administered 1 or more vaccines
at 360 veterinary hospitals from January 1, 2002, to December 31, 2003.
1 for remainder of key.
Breed Dogs vaccinated (n) VAAE (n) VAAE rate/10,000 dogs 95% CL
Dachshund 41,323 503 121.7 (111.4, 132.8)
Pug 20,214 188 93.0 (80.2, 107.2)
Boston Terrier 9,541 80 83.8 (66.5, 104.3)
Miniature Pinscher 15,310 117 76.4 (63.2, 91.5)
Chihuahua 68,839 524 76.1 (69.8, 82.9)
Maltese 20,663 138 66.8 (56.1, 78.9)
Miniature Schnauzer 15,296 99 64.7 (52.6, 78.7)
Jack Russell Terrier 23,717 129 54.4 (45.4, 64.6)
Toy Poodle 12,311 61 49.5 (37.9, 63.6)
Yorkshire Terrier 33,563 159 47.3 (40.3, 55.3)
Boxer 31,141 142 45.6 (38.4, 53.7)
Pomeranian 27,543 125 45.4 (37.8, 54.0)
Pekingese 9,516 40 42.0 (30.0, 57.2)
Shih Tzu 47,964 199 41.5 (35.9, 47.6)
English Bulldog 5,470 22 40.2 (25.2, 60.8)
Lhasa Apso 15,386 59 38.3 (29.2, 49.4)
Weimaraner 5,393 20 37.1 (22.7, 57.2)
Beagle 34,872 126 36.1 (30.1, 43.0)
Bichon Frise 13,444 46 34.2 (25.1, 45.6)
American Eskimo Dog 5,829 22 33.7 (23.7, 57.1)
American Cocker Spaniel 20,795 70 33.6 (26.3, 42.5)
Shetland Sheepdog 9,891 33 33.4 (23.0, 46.8)
Shar Pei 7,337 24 32.7 (21.0, 48.6)
Miniature Poodle 7,207 23 31.9 (20.2, 47.8)
Golden Retriever 41,779 126 30.2 (25.1, 35.9)
Basset Hound 7,828 23 29.4 (18.6, 44.1)
Welsh Corgi 5,511 16 29.0 (16.6, 47.1)
Siberian Husky 6,362 17 26.7 (15.6, 42.7)
Great Dane 5,211 13 24.9 (13.3, 42.6)
West Highland White Terrier 6,742 16 23.7 (13.6, 38.5)
Labrador Retriever 132,222 312 23.6 (21.1, 26.4)
Doberman Pinscher 6,520 15 23.0 (12.9, 37.9)
American Pit Bull Terrier 6,718 15 22.3 (12.5, 36.8)
Akita 6,161 13 21.1 (11.2, 36.1)
Mixed 44,188 89 20.1 (16.2, 24.8)
Australian Shepherd 16,221 30 18.5 (12.5, 26.4)
Dalmatian 7,234 13 18.0 (9.6, 30.7)
Australian Cattle Dog 5,702 10 17.5 (8.4, 32.2)
Border Collie 13,524 22 16.3 (10.2, 24.6)
Collie 5,708 9 15.8 (7.2, 29.9)
Chow Chow 23,387 32 13.4 (9.4, 19.3)
German Shepherd Dog 60,017 78 13.0 (10.3, 16.2)
Rottweiler 38,538 50 13.0 (9.6, 17.1)
Table 1—Incidence rate per 10,000 dogs and 95% confidence limits (CL) by breed for vaccine-associated
adverse events (VAAEs) diagnosed within 3 days of vaccine administration at 360 veterinary hospitals
from January 1, 2002, to December 31, 2003. Breeds listed if represented by ≥ 5,000 vaccinated dogs.
05-02-0089.qxp 9/13/2005 10:38 AM Page 1104
JAVMA, Vol 227, No. 7, October 1, 2005 Scientific Reports: Original Study 1105
phylaxis, 32 (0.7%) as urticaria, and 5 (0.1%) as cardiac
arrest. Death was reported in association with vaccina-
tion in 3 dogs (0.024 deaths/10,000 dogs vaccinated
[2.4 deaths/1,000,000]; 95% CI, 0.005 to 0.072).
Bivariate analyses—The population included
252,434 (20.6%) sexually intact males, 191,601
(15.6%) sexually intact females, 378,224 (30.8%)
neutered males, and 403,900 (32.9%) spayed females.
Among those groups, unadjusted VAAE rates/10,000
dogs were 32.9 (95% CI, 30.7 to 35.2), 35.7 (95% CI,
33.1 to 38.5), 39.1 (95% CI, 37.1 to 41.1), and 41.2
(95% CI, 39.8 to 43.8), respectively. Among all dogs,
the VAAE rates in 2002 (1,942/515,447 [37.7/10,000])
and 2003 (2,736/710,712 [38.5/10,000]) were not sig-
nificantly different (P = 0.47).
The VAAE rates decreased significantly as body
weight increased (P for trend < 0.001; Figure 1). For
all vaccines or for rabies vaccine alone, the VAAE rate
for 10.1- to 45.0-kg (22.2- to 99.0-lb) dogs was
approximately half the rate for dogs that weighed 0 to
10.0 kg (0 to 22.0 lb; P < 0.001; Figure 2). For rabies
vaccine administered alone, VAAE rates/10,000 dogs
that weighed 0 to 10.0 kg, 10.1 to 45.0 kg, and > 45 kg
were 32.1 (222/69,178), 15.3 (69/45,088), and 0.0
(0/1,966), respectively. In 586,817 dogs ≤ 9 months
old, the VAAE rate was 38.6/10,000 dogs vaccinated
(95% CI, 37.1 to 40.3), and the VAAE rate significant-
ly increased with age until 1.5 to 2.5 years of age
(VAAE rate, 53.8; 95% CI, 49.5 to 58.4; Figure 3). The
VAAE rates decreased progressively thereafter in older
age categories. Among breeds with 5,000 or more dogs
vaccinated, Dachshund, Pug, Boston Terrier, Miniature
Pinscher, and Chihuahua breeds had the highest rates
of VAAEs with 121.7, 93.0, 83.8, 76.4, and 76.1
adverse events/10,000 dogs vaccinated, respectively
(Table 1). The VAAE rate for mixed-breed dogs was in
the bottom quintile of all rates.
The risk of a VAAE significantly increased as the
number of vaccines administered per office visit
increased (P for trend < 0.001). Unadjusted
VAAE rates increased from 25.2 for a single
vaccine to 56.3/10,000 dogs when 6 vac-
cines were simultaneously administered
(705/279,330 and 397/70,554, respective-
ly). A strong linear dose-response relation-
= 0.985) was detected between
adjusted VAAE rates and the number of
vaccines administered. In all dogs, each
additional vaccine administered per office
visit increased the rate of a VAAE by 24.2%;
the rate increase was significantly (P <
0.001) greater in dogs that weighed 0 to
10.0 kg, compared with dogs that weighed
10.1 to 45.0 kg (27.3% vs 11.5%, respec-
tively; Figure 4). The 3 dogs with recorded
deaths each had received ≥ 4 vaccines at
their last office visit.
The VAAE rates associated with admin-
istration of a single dose of different vaccines
differed significantly (P < 0.001). The lowest
rate was observed with parenteral adminis-
tration of Bordetella vaccine (15.4/10,000; 82
VAAEs/53,238 doses), and the highest rate was observed
with Borrelia (Lyme disease) vaccine (43.7/10,000; 132
VAAEs/30,201 doses). The VAAE rates for Giardia vac-
cine (23.4/10,000; 97 VAAEs/41,447 doses), rabies vac-
cine (24.7/10,000; 293 VAAEs/118,765 doses), coron-
avirus vaccine (26.2/10,000; 15 VAAEs/5,735 doses), and
tospirosis vaccine (28.8/10,000; 86 VAAEs/29,852 doses)
were not significantly different (P = 0.526). There were
< 100 doses of parvovirus vaccine administered alone,
and no adverse events were recorded. There were 8
paired combinations of vaccines administered to at least
5,000 dogs. The VAAE rates for those combinations
ranged from 15.5/10,000 for Bordetella and Giardia vac-
cines (13 VAAEs/8,405 dogs) to 54.1/10,000 for concur-
rent administration of Borrelia and rabies vaccine (33
VAAEs/6,097 dogs). Concurrent administration of a
rabies and multivalent distemper vaccine to 25,171 dogs
resulted in 99 VAAEs or a VAAE rate of 39.3/10,000 dogs.
Multivariate analysis—A multivariate logistic
regression model including sex, neuter status, age,
weight, and number of vaccines received satisfied
requirements for goodness of fit (P = 0.72). In the final
model, the OR of a VAAE increased significantly (P <
0.001) as weight decreased and as the number of vac-
cines increased (Table 2). Risk for dogs that weighed ≤ 5
kg was more than 4 times the risk for dogs that weighed
> 45 kg (OR, 4.46; 95% CI, 2.67 to 7.46; P < 0.001).
Compared with the risk of a VAAE associated with a sin-
gle vaccination, simultaneous administration of 3 vac-
cines increased the risk approximately 50% (OR, 1.51;
95% CI, 1.37 to 1.67; P < 0.001), and with 5 simultane-
ous vaccinations, risk doubled (OR, 2.06; 95% CI, 1.82
to 2.33; P < 0.001). Risk of a VAAE was significantly
increased for neutered dogs, compared with sexually
intact dogs. The VAAE risk was greatest for dogs approx-
imately 1 to 3 years old and least for dogs ≥ 6 years of age.
Description of VAAEs—In a random sample of
400 dogs with a VAAE diagnosed within 3 days of vac-
Figure 4—Mean ± SEM VAAE rates by number of vaccines administered per
office visit at 360 veterinary hospitals in dogs of various BWs from January 1,
2002 to December 31, 2003.
Figure 1 for remainder of key.
05-02-0089.qxp 9/13/2005 10:38 AM Page 1105
1106 Scientific Reports: Original Study JAVMA, Vol 227, No. 7, October 1, 2005
cine administration, the predominant clinical signs
were facial or periorbital edema (30.8% [123/400]),
wheals or urticaria (20.8% [83/400]), generalized pru-
ritus (15.3% [61/400]), and vomiting (10.3%
[41/400]). Localized vaccination-site reactions (eg,
swelling, inflammation, or soreness) represented 8.0%
(32/400) of VAAEs, and systemic nonspecific signs (eg,
fever, lethargy, or anorexia) represented 5.5% (22/400)
of VAAEs. Collapse was the only clinical sign recorded
in 4 (1.0%) dogs. Most dogs in this sample were treat-
ed with an antihistamine and glucocorticoid (34.5%
[138/400]), an antihistamine alone (22.5% [90/400]),
or glucocorticoid alone (11.5% [46/400]). Less com-
monly used treatments included fluids, oxygen, diuret-
ics, or nonsteroidal anti-inflammatory drugs; epineph-
rine was administered to 3.2% (13/400) of dogs. For 69
of 400 (17.0%) dogs, no treatment was prescribed.
By use of this large population of dogs that
received primary care, it was possible to accurately
estimate the incidence rate of practitioner-diagnosed
VAAEs that occurred within 3 days of vaccine adminis-
tration. Although most VAAEs were recorded the same
day as vaccination, the 3-day period was selected as a
reasonable time for owners to report events or return
their dog to a veterinarian for examination. This peri-
od increased the likelihood that the observed event was
the result of a vaccination but excluded sequelae with
a longer latency period. The use of a primary care prac-
tice database permitted inclusion of VAAEs that may
not have been reported to industry or a federal agency
via a separate passive system. Rates for VAAEs were
determined by use of the actual number of vaccine
doses administered; therefore, a practice database pro-
vided more comprehensive event (numerator) data
while also providing population (denominator) data.
Traditional passive surveillance systems that receive
and summarize adverse event reports alone are not
useful for determination of incidence rates or potential
risk factors and are often characterized by under-
reporting of events.
National vaccine sales data have been used as the
denominator in some calculations of VAAE rates, but
such calculations result in estimates per dose of vaccine
sold rather than per dog, cannot determine the effect of
concurrently administered multiple vaccines, and are
also influenced by underreporting of adverse events.
Such estimates have reported
VAAE rates ranging
from 0.13 to 0.40/10,000 vaccine doses, whereas preli-
censure clinical studies reported postvaccination reac-
tions in excess of 1% or > 100 reactions/10,000 vaccine
In the present study, a VAAE rate of 0.38%
(approx 38 adverse events/10,000 vaccinated dogs or
13/10,000 doses administered) was found.
The risk of a VAAE in this study population was
inversely related to a dog’s weight. This weight-
response relationship was previously suggested by
results of a study
in which dogs of toy breeds had sig-
nificantly more suspected VAAEs than other dogs,
although body weight was not evaluated. The manufac-
turers’ recommended dose for all vaccines administered
in our study was 1 mL regardless of body weight, and
all vaccines were from single-dose vials. Vaccines, in
contrast to virtually all veterinary pharmaceuticals, are
prescribed on a 1-dose-fits-all basis, rather than by body
weight. Prelicensing clinical trials investigate the safety
of vaccines with doses in excess of label directions but
only in a limited number of dogs. The results of this
study suggest that trials in dogs that weigh > 10 kg
underestimate the expected VAAE rate in smaller dogs.
Prelicensing clinical trials also investigate the safe-
ty of vaccines in several hundred dogs at multiple hos-
pital locations, but specific breeds may be under- or
overrepresented. Mature weights of dogs of different
breeds may vary by 5 to 10 times and occasionally by
> 50 times. Therefore, a 1-mL vaccine dose results in a
ratio of vaccine volume received per kilogram of body
weight that can vary widely. When multiple vaccines
are simultaneously administered to a dog, the ratio of
volume received per kilogram of body weight per
patient also varies. The importance of this volume-to-
weight ratio in relation to adverse event risk was evi-
dent in this study by the increase in VAAE rates as the
number of simultaneously administered vaccine doses
increased, even when adjusted for weight.
Factors known to cause vaccine reactions include
the primary vaccine agent or antigen, adjuvants,
preservatives, stabilizers, and residues from tissue cul-
tures used in vaccine production.
The overall for-
Risk factor Odds ratio 95% CL
Sex and neuter
Male, sexually 1.00 NA NA
Female, sexually 1.06 (0.95, 1.18) 0.287
Male, neutered 1.27 (1.16, 1.38) ⬍ 0.001
Female, spayed 1.38 (1.27, 1.51) ⬍ 0.001
No. of vaccines per
1 1.00 NA NA
2 1.36 (1.23, 1.51) ⬍ 0.001
3 1.51 (1.37, 1.67) ⬍ 0.001
4 1.91 (1.73, 2.11) ⬍ 0.001
5 2.06 (1.82, 2.33) ⬍ 0.001
6 2.31 (2.03, 2.62) ⬍ 0.001
0 to 5 kg 4.46 (2.67, 7.46) ⬍ 0.001
⬎ 5 to 10 kg 4.21 (2.51, 7.05) ⬍ 0.001
⬎ 10 to 15 kg 3.00 (1.77, 5.06) ⬍ 0.001
⬎ 15 to 20 kg 2.52 (1.48, 4.29) 0.001
⬎ 20 to 25 kg 1.93 (1.13, 3.32) 0.017
⬎ 25 to 30 kg 1.93 (1.12, 3.33) 0.018
⬎ 30 to 35 kg 1.91 (1.10, 3.32) 0.022
⬎ 35 to 40 kg 1.68 (0.94, 3.00) 0.081
⬎ 40 to 45 kg 1.54 (0.81, 2.92) 0.183
⬎ 45 kg 1.00 NA NA
2 to 9 mo 1.00 NA NA
1 y 1.51 (1.37, 1.67) ⬍ 0.001
2 y 1.64 (1.49, 1.81) ⬍ 0.001
3 y 1.35 (1.20, 1.52) ⬍ 0.001
4 to 5 y 1.06 (0.95, 1.19) 0.282
6 to 8 y 0.79 (0.70, 0.90) ⬍ 0.001
⬎ 8 y 0.50 (0.43, 0.59) ⬍ 0.001
NA = Not applicable.
To convert kilograms to pounds, multiply by 2.2.
Table 2—Adjusted odds ratios and 95% CLs (multivariate logistic
regression) for potential risk factors for VAAEs in 1,226,159 dogs.
05-02-0089.qxp 9/13/2005 10:38 AM Page 1106
JAVMA, Vol 227, No. 7, October 1, 2005 Scientific Reports: Original Study 1107
mulation of various vaccine components (eg, antigen,
adjuvants, and diluent) is proprietary information that
was unavailable for analysis in our study; thus, the
variation in VAAE rates among single-antigen vaccines
may not be solely attributable to the primary vaccine
antigen. The nearly linear dose-response relationship
between number of vaccines simultaneously adminis-
tered and the VAAE rate suggests that vaccine compo-
nents other than the primary antigen may contribute to
adverse events. In a recent study
of 8 dogs that devel-
oped immediate-type allergic reactions and had high
concentrations of specific serum IgE against the vac-
cines, 7 had specific IgE against fetal calf serum.
The increased risk of VAAEs in smaller breeds in
this study was consistent with the weight-related find-
ings; however, a genetic predisposition of some breeds to
VAAEs is also possible. Although Dachshunds have been
considered at increased risk for VAAEs,
have not found a difference in hypersensitivity reactions
for this breed, compared with all other breeds. Genetic
susceptibility to allergy occurs in humans,
family and breed genetics may play a similar role in dogs
with respect to VAAEs. Likewise, because of genetic het-
erogeneity, the relatively low VAAE rate observed in
mixed-breed dogs suggests that laboratory safety trials
that use such dogs may underestimate the VAAE rates
that would occur in purebred dogs. This is important
because purebred dogs comprise at least two thirds of
the US dog population.
The risk of allergic reaction has been reported
increase after the third or fourth injection of a vaccine
(ie, a booster response). In our study, VAAE risk
increased for dogs up to 2 years of age and then
declined thereafter. In a controlled study,
trations were found to be greatest after vaccination at 2
and 3 years of age, corresponding to a dog’s third and
fourth doses of vaccine, but were not as increased after
vaccination at 4 years of age. Because of variations in
canine vaccination protocols, the third injection of vac-
cine may constitute the last puppy vaccination or a
booster at 1 year of age. The decline in the VAAE rate
observed after 2 years of age in this study may have
been attributable to allergen desensitization, initiation
of alternative vaccination protocols in predisposed
dogs, or owner refusal to revaccinate dogs that previ-
ously had a VAAE.
Neutering appeared to increase risk of a VAAE
more than sex. Females mount stronger immune
responses after vaccination or infection than males
because of a dimorphic enhancing effect of estrogens
and a protective effect of androgens.
reduces serum estrogen and testosterone concentra-
tions and also removes their negative feedback on the
pituitary gland, resulting in significant increases in fol-
licle-stimulating hormone and luteinizing hormone
concentrations in female and male dogs.
itary hormones may independently, or through interac-
tion with primary sex hormones, influence the
immune response to vaccination.
In this study, rates for adverse events associated
with the administration of single vaccines were not sig-
nificantly different for multivalent distemper (with 4
Leptospira serovars), Giardia, rabies, and coronavirus
vaccines. Vaccines containing inactivated Leptospira
bacterins have been considered to be more allergenic
than tissue culture lines of virus vaccines, but newer
subunit vaccines have been developed to reduce this
The multivalent vaccine used by the veteri-
nary hospitals in our study was a purified Leptospira
product that contained the immunogenic envelope,
and increased allergenicity of vaccines containing
Leptospira was not clearly detected in this large popu-
lation of dogs. The risk of hypersensitivity reactions
after administration of Borrelia vaccine has been con-
These reactions can be attributable
to an immunologic response to proinflammatory sur-
which is a possible cause for the
increased VAAE rates associated with administration of
the vaccine in this study. Event rates for specific vac-
cines in this study may or may not be representative of
other vaccines, but data pertaining to other products
are lacking for comparison.
Clinical signs of VAAEs and the predominance of
same-day events in this study were generally indica-
tive of immediate-type hypersensitivity reactions.
Signs associated with immediate hypersensitivity vary
by species and are related to the location of mast cells
that degranulate in the presence of an allergen.
Cutaneous or dermatologic events were most com-
monly reported by veterinarians within 3 days of vac-
cine administration, consistent with most reports of
suspected canine VAAEs submitted to the US
Adverse events documented in the
patient medical notes reviewed in this study were not
generally described as life-threatening, even when the
diagnosis code was anaphylaxis.
The size of large practice databases can make full
validation of diagnoses or a complete record review of
all patient information exorbitantly time-consuming.
Validation of a subset of records within large databases
has been recommended
to overcome this problem,
and medical notes were reviewed for a subset of the
population in our study. In an assessment of the impact
of validating only 1% to 2% of a study population
against paper-based medical records, errors of misclas-
sified diagnoses only modestly bias outcome rate ratios
toward the null hypothesis.
The use of practice data-
bases also presents new but surmountable challenges
in processing and analyzing extremely large data sets,
sometimes exceeding 500 MB in storage size.
Adverse events in this study, as in all postmarketing
surveillance systems, were based on diagnoses made by
different practitioners. Although diagnoses consistent
with vaccine reactions were selected by practitioners
from the available codes in the software, computerized
databases are dependent on coded outcomes and some
codes may be nonspecific.
Standardized case defini-
tions for VAAEs are not available in veterinary medicine
and are only currently being addressed in human med-
Bias may therefore be introduced into a study if
adverse events are not captured by the existing data-
base, either because they do not result in an office con-
sultation or because appropriate coding is not chosen.
Biases that potentially affect studies with large data-
bases are not unlike those of nonautomated data
sources that require case definitions, manual abstrac-
05-02-0089.qxp 9/13/2005 10:38 AM Page 1107
1108 Scientific Reports: Original Study JAVMA, Vol 227, No. 7, October 1, 2005
tion, coding, and computerization. In a large practice,
reporting biases or misdiagnoses that may potentially
be introduced by a few individuals are reduced because
of the large number of hospitals contributing data. The
random effects variation related to individual veterinar-
ians was not evaluated in the multivariate model, but
incorporation of dog random effects into analysis
resulted in minimal change in ORs because of the large
number of dogs (data not shown).
The population size in large databases produces
statistical power such that calculated P values are often
small and within traditional significance levels. Study
results must therefore not be interpreted solely on sig-
nificance but rather on clinical importance. Although
VAAE rates in this study were higher than estimated
rates reported from passive surveillance systems, these
events are relatively infrequent or uncommon in a hos-
pital population. Dog characteristics that increase
VAAE risk should be highlighted in risk communica-
tion with clients, but risk aversion of adverse events
must be tempered with the risk of sickness and death
from infectious diseases.
Research will be required to characterize the pri-
mary allergenic components of different licensed vet-
erinary vaccines, and then it will need to be deter-
mined whether vaccine allergenicity and volume can
be reduced while immunologic protection is main-
tained, particularly for smaller dogs. Until such time,
practitioners may choose to reduce the number of vac-
cines simultaneously administered to small dogs (but
not reduce the volume of an individual vaccine) and
alert owners to risk factors for VAAEs.
Although VAAE rates were higher for certain risk
factors or some vaccines, compared with other factors
or vaccines, the rates for VAAEs were low in the overall
population. Evidence of VAAEs does not indicate that
vaccines are not safe but rather that there is a small risk
of adverse events associated with certain dog factors or
vaccines. Premarketing safety studies, when fiscally or
logistically limited in size, will remain limited in power
to detect rare adverse events that may be more common
among animals with particular risk factors.
a. PetWare, Banfield, The Pet Hospital, Portland, Ore.
b. Brochicine, Biocor Animal Health Inc, Omaha, Neb.
c. Duramune Cv-K, Fort Dodge Animal Health, Fort Dodge, Iowa.
d. Duramune Max 5/4L, Fort Dodge Animal Health, Fort Dodge,
e. Giardia Vax, Fort Dodge Animal Health, Fort Dodge, Iowa.
f. Lymevax, Fort Dodge Animal Health, Fort Dodge, Iowa.
g. Duramune Max Pv, Fort Dodge Animal Health, Fort Dodge,
h. Rabvac 3, Fort Dodge Animal Health, Fort Dodge, Iowa.
i. STATA, version 8.2, StataCorp, College Station, Tex.
1. Gaskell RM, Gettinby G, Graham SJ, et al. Veterinary
Products Committee working group report on feline and canine vac-
cination. Vet Rec 2002;150:126–134.
2. Gobar GM, Kass PH. World Wide Web-based survey of vac-
cination practices, postvaccinal reactions, and vaccine site-associat-
ed sarcomas in cats. J Am Vet Med Assoc 2002;220:1477–1482.
3. Rose NR. Immunologic hazards associated with vaccination
of humans. J Autoimmun 2000;14:11–13.
4. Rundle R. Are annual shots overkill? The Wall Street Journal
5. Varricchio F, Iskander J, Destefano F, et al. Understanding
vaccine safety information from the Vaccine Adverse Event
Reporting System. Pediatr Infect Dis J 2004;23:287–294.
6. Meyer EK. Vaccine-associated adverse events. Vet Clin North
Am Small Anim Pract 2001;31:493–514.
7. Chen RT, DeStefano F, Davis RL, et al. The Vaccine Safety
Datalink: immunization research in health maintenance organiza-
tions in the USA. Bull World Health Organ 2000;78:186–194.
8. Verstraeten T, DeStefano F, Chen RT, et al. Vaccine safety
surveillance using large linked databases: opportunities, hazards and
proposed guidelines. Expert Rev Vaccines 2003;2:21–29.
9. DeStefano F. The Vaccine Safety Datalink project.
Pharmacoepidemiol Drug Saf 2001;10:403–406.
10. Ellenberg SS, Braun MM. Monitoring the safety of vaccines:
assessing the risks. Drug Saf 2002;25:145–152.
11. Chen RT, Rastogi SC, Mullen JR, et al. The Vaccine Adverse
Event Reporting System (VAERS). Vaccine 1994;12:542–550.
12. Brooks R. Adverse reactions to canine and feline vaccines.
Aust Vet J 1991;68:342–344.
13. Gray AK. Cat and dog vaccination: results from the Suspected
Adverse Reaction Surveillance Scheme. Vet Rec 1998;143:455.
14. Georgitis JW, Fasano MB. Allergenic components of vaccines
and avoidance of vaccination-related adverse events. Curr Allergy Rep
15. Roth JA. Mechanistic bases for adverse vaccine reactions
and vaccine failures. Adv Vet Med 1999;41:681–700.
16. Martinod S. Adverse effects of vaccination. In: Pastoret P,
Blancou J, Vannier P, eds. Veterinary vaccinology. Amsterdam:
Elsevier Science, 1997;574–580.
17. HogenEsch H, Azcona-Olivera J, Scott-Moncrieff C, et al.
Vaccine-induced autoimmunity in the dog. Adv Vet Med 1999;41:
18. Ohmori K, Masuda K, Maeda S, et al. IgE reactivity to vaccine
components in dogs that developed immediate-type allergic reactions
after vaccination. Vet Immunol Immunopathol 2005;104:249–256.
19. Greene CE. Immunoprophylaxis and immunotherapy. In:
Greene CE, ed. Infectious diseases of the dog and cat. 2nd ed.
Philadelphia: WB Saunders Co, 1998;717–750.
20. Huang SK, Mathias RA, Ehrlich E, et al. Evidence for asth-
ma susceptibility genes on chromosome 11 in an African-American
population. Hum Genet 2003;113:71–75.
21. Ober C. Susceptibility genes in asthma and allergy. Curr
Allergy Asthma Rep 2001;1:174–179.
22. American Pet Products Manufacturers Association.
2003–2004 APPMA national pet owners survey. Greenwich, Conn:
American Pet Products Manufacturers Association, 2003;7.
23. HogenEsch H, Dunham AD, Scott-Moncrieff C, et al. Effect
of vaccination on serum concentrations of total and antigen-specific
immunoglobulin E in dogs. Am J Vet Res 2002;63:611–616.
24. Beagley KW, Gockel CM. Regulation of innate and adaptive
immunity by the female sex hormones oestradiol and progesterone.
FEMS Immunol Med Microbiol 2003;38:13–22.
25. Verthelyi D. Sex hormones as immunomodulators in health
and disease. Int Immunopharmacol 2001;1:983–993.
26. Olson PN, Mulnix JA, Nett TM. Concentrations of luteiniz-
ing hormone and follicle-stimulating hormone in the serum of sexu-
ally intact and neutered dogs. Am J Vet Res 1992;53:762–766.
27. Klingborg DJ, Hustead DR, Curry-Galvin EA, et al. AVMA
Council on Biologic and Therapeutic Agents’ report on cat and dog
vaccines. J Am Vet Med Assoc 2002;221:1401–1407.
28. Tizard I. Type I hypersensitivity. Veterinary immunology: an
introduction. 7th ed. Philadelphia: WB Saunders Co, 2004;308–323.
29. Mullooly J, Drew L, DeStefano F, et al. Quality of HMO vac-
cination databases used to monitor childhood vaccine safety. Vaccine
Safety DataLink Team. Am J Epidemiol 1999;149:186–194.
05-02-0089.qxp 9/13/2005 10:38 AM Page 1108