Content uploaded by Rowena Mary Anne Packer
Author content
All content in this area was uploaded by Rowena Mary Anne Packer on Oct 30, 2015
Content may be subject to copyright.
RESEARCH ARTICLE
Impact of Facial Conformation on Canine
Health: Brachycephalic Obstructive Airway
Syndrome
Rowena M. A. Packer
1
*, Anke Hendricks
1
, Michael S. Tivers
2
, Charlotte C. Burn
3
1Department of Clinical Science and Services, Royal Veterinary College, Hatfield, Hertfordshire, United
Kingdom, 2School of Veterinary Sciences, University of Bristol, Langford House, Langford, Bristol, BS40
5DU, United Kingdom, 3Department of Production and Population Health, Royal Veterinary College,
Hatfield, Hertfordshire, United Kingdom
*rpacker@rvc.ac.uk
Abstract
The domestic dog may be the most morphologically diverse terrestrial mammalian species
known to man; pedigree dogs are artificially selected for extreme aesthetics dictated by for-
mal Breed Standards, and breed-related disorders linked to conformation are ubiquitous
and diverse. Brachycephaly–foreshortening of the facial skeleton–is a discrete mutation
that has been selected for in many popular dog breeds e.g. the Bulldog, Pug, and French
Bulldog. A chronic, debilitating respiratory syndrome, whereby soft tissue blocks the air-
ways, predominantly affects dogs with this conformation, and thus is labelled Brachyce-
phalic Obstructive Airway Syndrome (BOAS). Despite the name of the syndrome, scientific
evidence quantitatively linking brachycephaly with BOAS is lacking, but it could aid efforts
to select for healthier conformations. Here we show, in (1) an exploratory study of 700 dogs
of diverse breeds and conformations, and (2) a confirmatory study of 154 brachycephalic
dogs, that BOAS risk increases sharply in a non-linear manner as relative muzzle length
shortens. BOAS only occurred in dogs whose muzzles comprised less than half their cranial
lengths. Thicker neck girths also increased BOAS risk in both populations: a risk factor for
human sleep apnoea and not previously realised in dogs; and obesity was found to further
increase BOAS risk. This study provides evidence that breeding for brachycephaly leads to
an increased risk of BOAS in dogs, with risk increasing as the morphology becomes more
exaggerated. As such, dog breeders and buyers should be aware of this risk when selecting
dogs, and breeding organisations should actively discourage exaggeration of this high-risk
conformation in breed standards and the show ring.
Introduction
The domestic dog is the most morphologically diverse terrestrial mammalian species known to
man [1]; however, pedigree dogs are strongly selected for aesthetics dictated by formal Breed
Standards, and breed-related disorders linked to conformation are ubiquitous and diverse
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 1/21
OPEN ACCESS
Citation: Packer RMA, Hendricks A, Tivers MS, Burn
CC (2015) Impact of Facial Conformation on Canine
Health: Brachycephalic Obstructive Airway
Syndrome. PLoS ONE 10(10): e0137496.
doi:10.1371/journal.pone.0137496
Editor: Carlos E. Ambrósio, Faculty of Animal
Sciences and Food Engineering, University of São
Paulo, BRAZIL
Received: October 10, 2014
Accepted: August 18, 2015
Published: October 28, 2015
Copyright: © 2015 Packer et al. This is an open
access article distributed under the terms of the
Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any
medium, provided the original author and source are
credited.
Data Availability Statement: All relevant data are
within the paper and its supporting information files.
Funding: This study was funded by UFAW (www.
ufaw.org.uk) and Dogs Trust (www.dogstrust.org.uk)
while the primary author was supported by a joint
BBSRC/RVC doctoral studentship (www.bbsrc.ac.uk;
www.rvc.ac.uk). The funders had no role in study
design, data collection and analysis, decision to
publish, or preparation of the manuscript.
Competing Interests: Michael S. Tivers was
employed by Cave Veterinary Specialists during part
[2,3]. Artificial selection for desired traits in domestic animals can cause unintended changes
in other traits, which can be pathological, e.g. lameness in highly productive dairy cows [4] and
broiler chickens [5]. The mechanisms underlying such pathologies include the inadvertent
genetic consequences of inbreeding or linkage disequilibrium [3], or may be direct physical
consequences of the desired trait [2]. Brachycephaly, or foreshortening of the facial skeleton, is
a discrete skeletal mutation [6] where altered growth of the basioccipital and basisphenoid
bones manifests in a shortening of the basicranial axis [7]. This results in the characteristic
short-muzzled, or flat facial conformation that has been intensely selected for by dog breeders
to develop many popular companion dog breeds, due to the anthropocentric appeal of their
juvenile, human like features [8]. Despite the increasing popularity of brachycephalic breeds
such as the Pug, Bulldog and French Bulldog in the UK and internationally [9], this conforma-
tion is not benign and is associated with several inherited disorders of the head and neck
[2,10].
Anatomical abnormalities
Brachycephalic Obstructive Airway Syndrome (BOAS) is a debilitating respiratory syndrome
that predominantly affects brachycephalic dogs, whereby soft tissue blocks the airways during
respiration (Video A in S3 File). A previous study of this disorder dichotomised breeds as
brachycephalic or not, and showed that 39 of their 45 BOAS cases occurred in brachycephalic
breeds, with brachycephaly conferring an odds ratio of 38 on the risk of BOAS [11]. The condi-
tion arises in brachycephalic animals because, despite a marked reduction in facial skeleton
length [7], the soft tissue structures of the oral cavity (e.g. soft palate, tongue, tonsils) are not
proportionally reduced [12]. As an affected dog matures, the compacted soft tissue increasingly
impede airflow by blocking the larynx and nasopharynx, and impairs the thermoregulatory
function of the nose via internal [13] and external nasal obstruction [14]. Within the nose, tur-
binate growth in young brachycephalic dogs continues despite inhibition of growth of the mid-
face, resulting in relatively oversized turbinates [15]. The resulting contact between turbinate
lamella mucosal surfaces impedes nasal airflow. Externally, the wing of the nostril (ala nasi) is
congenitally deformed in many brachycephalic dogs [14], with a narrowing of the nostrils (‘ste-
notic nares’). These primary abnormalities can result in markedly increased respiratory efforts
to overcome airway resistance, fostering collapse of the airway [16], most commonly the larynx
[17,18]. Laryngeal collapse is the most common and serious secondary change associated with
BOAS, with a guarded prognosis for late stage laryngeal collapse [19,20]. The functional prob-
lems subsumed under the term BOAS are thus the result of the skeletal shortening and the rela-
tionship between the facial skeleton and the soft and hard tissue structures contained within it.
Dogs of breeds not traditionally classified as brachycephalic have rarely been diagnosed with
BOAS in previous studies, such as the Chow Chow, Rottweiler and Pomeranian [21]. These
breeds may be classed as mesocephalic (a head of medium proportions); however, without
morphometric data it is not possible to ascertain whether these individuals were short-muzzled
for their breed, or whether other risk-factors for BOAS were present.
Clinical signs and welfare impact
BOAS is characterised by a chronic shortness of breath and subsequent difficulties in exercising
(e.g. walking, running and playing), a propensity to overheating, increased and abnormal respi-
ratory noise (e.g. snoring and snorting), low blood oxygen levels and consequently, collapse
[22]. BOAS affected dogs are prone to heat stroke, which can result in death [23]. Severely
affected individuals exhibit laboured breathing, often adopting a wide stance with their elbows
abducted from their chest, with the use of accessory abdominal musculature [24] and over
Impact of Facial Conformation on Canine Health: Airway Disease
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 2/21
of this study. There are no patents, products in
development or marketed products to declare. This
does not alter the authors’adherence to all the PLOS
ONE policies on sharing data and materials, as
detailed online in the guide for authors.
inflation of the chest [25] observed. BOAS has been a recognised disorder for many years, with
surgical techniques developed to treat this syndrome described as early as the 1940’s[26,27].
BOAS has potentially severe welfare consequences [28], with the most affected dogs
described as having “little or no activity”because they are fully occupied just with breathing
[29]. Any form of stress, exercise or excitement can cause severe respiratory distress in such
dogs, and occasionally even death [16,17] Minor aggravations can lead to severe respiratory
distress [23], with arousal by both negative and positive experiences (e.g. stress, but also exer-
cise and excitement) acting as aggravators [23,25]. Clinical signs of BOAS can be evident whilst
the dog is awake or asleep, with audible snoring reported in 100% of BOAS affected dogs vs.
21% of unaffected dogs in a recent study [22], and sleep-disordered breathing (including epi-
sodes of ‘apnoea’, cessation of breathing) well researched in the Bulldog [30]. The effects of
BOAS are not just limited to the respiratory system, with chronic negative pressure in the chest
cavity leading to gastrointestinal tract lesions, manifested as clinical signs such as gagging,
regurgitation and vomiting [31]. Clinical signs are often severe by 12 months of age [32] and
are life-long thereafter.
Conformation and disease
Despite the name of the syndrome, scientific literature quantitatively linking brachycephaly
with BOAS is lacking, with this hypothesised relationship based on overrepresentation of
brachycephalic breeds in international case series of BOAS (e.g. [11]). BOAS has been reported
in over 10 brachycephalic breeds internationally [22]; however, brachycephaly is not a binary
trait. Instead, relative muzzle length is a highly variable aspect of canine skull morphology [33].
International inquiries regarding dog breeding practices [34,35] have called for information
quantifying the extent to which BOAS relates to craniofacial conformation, and therefore what
constitutes a muzzle being ‘too short’, which was previously unknown. These quantitative lim-
its were initially proposed by the Council of Europe [35], where it was suggested that “Maxi-
mum values for the shortness of skull,respectively the nose to avoid breathing difficulties”should
be set.
Breed standards
In recent surveys of veterinarians [36] and other canine stakeholders in the UK [37], altering
breed standards was the most common suggestion to reduce the prevalence of inherited dis-
eases in pedigree dogs, with veterinarians also strongly disagreeing with the statement that
breed standards support the health and welfare of dogs [36]. Data linking inherited diseases
with morphologies encouraged by breed standards could thus encourage the revision of breed
standards internationally, with quantative limits included to encourage the necessary changes
and reverse the current trend towards ‘short’muzzles, as evidenced in current breed standards
(Table 1), if this morphology does indeed increase BOAS risk.
Aims
The aim of this study was to confirm and quantify the extent to which the brachycephalic cra-
niofacial conformation is associated with BOAS risk, and whether more extreme brachyce-
phalic conformations are at a higher risk of BOAS than moderate or mildly brachycephalic
morphologies. This is with the practical aim of being able to provide data upon which to base
quantitative recommendations regarding the degree of brachycephaly that is ‘acceptable’to
breed for, helping reduce risk of this disorder. This study deliberately used a morphometric
methodology that is non-invasive and easily applicable to conscious dogs without the need for
special equipment. In addition, we aimed to investigate whether BOAS is related to craniofacial
Impact of Facial Conformation on Canine Health: Airway Disease
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 3/21
conformation per se or some other aspect of breed affiliation, e.g. genetic background or life-
style factors. We did this by first examining the relationship between conformation and BOAS
in all breeds and cross breeds, and then limiting the analysis solely to affected breeds, to see if
any relationship remained within affected breeds.
Study Plan
To investigate how craniofacial morphology related to respiratory function, we examined the
conformation and clinical status of two populations of dogs; firstly, dogs of any breed or cross-
breed entering a veterinary referral hospital for any condition (referred to as ‘Study 1’)togain
general estimates of the relationship between BOAS and muzzle length in a large, genetically and
morphometrically varied population of domestic dogs. This was followed by studying a second
population outside the referral hospital environment (referred to as ‘Study 2’), this time focussing
on brachycephalic dogs only, to test the robustness of the initial estimates generated in Study 1.
Materials and Methods- Study 1
Morphometrics
The conformation of all dogs was measured using established breed-defining measurement
protocols [38]. Thirteen conformational features shown to be breed-defining were
Table 1. Kennel Club and American Kennel Club breed standards of popular brachycephalic breeds,
describing ‘short’muzzles which may put them at risk of BOAS. The ‘nose’refers to the nose leather,
but can constitute a large proportion of the muzzle length in brachycephalic dogs.
Breed standard text referring to the relative length of the ‘muzzle’,‘face’and/or
‘nose’
Breed Kennel Club American Kennel Club
Pug Muzzle relatively short, blunt,
square [61]
The muzzle is short, blunt, square, but not upfaced
[64]
French
Bulldog
Nose relatively short [63] The muzzle broad, deep and well laid back. The stop
well defined, causing a hollow groove between the
eyes with heavy wrinkles forming a soft roll over the
extremely short nose [65]
Bulldog Face relatively short [62] The face, measured from the front of the cheekbone
to the tip of the nose, should be extremely short, the
muzzle being very short, broad, turned upward [66]
Pekingese Muzzle must be evident, but may
be relatively short and wide [75]
Muzzle—It is very flat, broad, and well filled-in below
the eyes. Nose—It is broad, short and black.[76]
Griffon
Bruxellois
Relatively short, wide muzzle [77] Nose very black, extremely short, its tip being set
back deeply between the eyes so as to form a lay-
back.[78]
Japanese
Chin
Muzzle short, wide [79] Muzzle—short and broad with well-cushioned cheeks
and rounded upper lips that cover the teeth. Nose—
very short.[80]
Boston
Terrier
Muzzle relatively short, square,
wide [81]
The muzzle is short, square, wide and deep and in
proportion to the skull. It is free from wrinkles, shorter
in length than in width or depth; not exceeding in
length approximately one-third of the length of the
skull. [82]
Shih Tzu Muzzle of ample width, square,
short [83]
Square, short, unwrinkled, with good cushioning, set
no lower than bottom eye rim; never downturned.
Ideally, no longer than 1 inch from tip of nose to stop,
although length may vary slightly in relation to overall
size of dog. [84]
doi:10.1371/journal.pone.0137496.t001
Impact of Facial Conformation on Canine Health: Airway Disease
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 4/21
measured: muzzle length, cranial length, head width, eye width, neck length, neck girth, chest
girth, chest width, body length, height at the withers and height at the base of tail, fore limb cir-
cumference and hind limb circumference (all in cm). All measurements were made to the near-
est millimetre. Muzzle length was defined as the distance (mm) from the dorsal tip of the nasal
planum to the stop, and was measured from the tip of the nose to just between the eyes where
the inside corners of the eyes meet (as the stop is not discernible on longer-muzzled dogs with
a less pronounced facial angle). Cranial length (CL) was defined as the distance (mm) from the
stop to the occipital protuberance, following the curve of the cranial surface (rather than a lin-
ear measure), and was measured from just between the eyes up the face, between the ears, to
the back of the head where the bony process projects out. Both measurements were taken using
a standard 1 m soft measuring tape. The degree of brachycephaly (facial foreshortening) was
quantified by the craniofacial ratio (CFR): the muzzle length divided by the cranial length [22]
(Fig 1). All dogs were examined for the presence of a nasal fold; defined as a discernible fold of
skin on the dorsal surface of the muzzle that was present without manipulation of the skin, and
could be easily grasped between vernier callipers; this was recorded as a binary trait.
The craniofacial morphometric parameters used in this study were not those previously cre-
ated for the examination of dry skulls using callipers to measure linear distances between set
points [39,40], or those used in CT/MRI analysis of the skull [41,42]. Instead, the measure-
ments used here were first described in a previous large-scale study (>1000 dogs) of a wide
variety of dog breeds in the United States [38], and created with the purpose of being easily rep-
licated by owners in the home; thus these aligned well with our aim of using measurements
that breeders and owners can take with simple equipment easily available to them. The cranio-
facial ratio previously devised by the authors of the current study [22] produces intuitive fig-
ures that reflect the degree of foreshortening of the muzzle. Further studies may focus on the
inter-rater and intra-rater reliability of this measure.
Weight (kg) was measured in all dogs on regularly calibrated digital scales. Body condition
score (BCS) was assessed on a a 9 point scale [43] by a single-rater (RMAP). To test for poten-
tial effects of other aspects of body shape and size, as seen in a previous study of conformation-
related disease [21], Principal Component Analysis of the remaining measurements was car-
ried out, to attempt to capture two overarching aspects of each dog’s morphology: overall
Fig 1. Diagram of how to measure (i) cranial length (A-B) and (ii) muzzle length (B-C). Measurements were taken using a soft measuring tape. Cranial
length is defined as the distance (mm) from the occipital protuberance (A) to the stop (B). Muzzle length is defined as the distance (mm) from the dorsal tip of
the nasal planum (C) to the stop (B). The precise locations of the nasal planum, stop, and occipital protuberance are determined through palpation as well as
visually, but the lettering indicates their approximate locations on the photographs. This is demonstrated in (left-right) an extremely brachycephalic Pug
(CFR = 0.08), a moderately brachycephalic Bulldog cross (CFR = 0.23) and a mildly brachycephalic Boxer (CFR = 0.35).
doi:10.1371/journal.pone.0137496.g001
Impact of Facial Conformation on Canine Health: Airway Disease
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 5/21
skeletal body size (Principal Component 1: PC1) and the ‘thickness/broadness’of the dog’s
body (Principal Component 2: PC2); components that were previously identified from a simi-
lar canine morphometric data set [38]. Low PC1 indicates a small skeletal size (e.g. a Chihua-
hua) while high PC1 indicates a large skeletal size (e.g. a Great Dane). Low PC2 indicates a
narrow and slim-boned body shape (e.g. a Greyhound), while a high PC2 indicates a broad and
thick-boned body shape (e.g. a Mastiff). Cranial length and muzzle length were omitted from
that analysis, so that these variables were not included in the statistical models twice. Principal
components were extracted based on eigenvalue, with only those greater than one extracted.
Recruitment of dogs
Between December 2010 and January 2012, every dog referred to the Royal Veterinary College
Small Animal Referral Hospital (RVC SARH) was considered for inclusion in Study 1. Owners
of dogs referred to any clinical service for a routine appointment were approached. All dogs
were considered for recruitment prior to their arrival at the hospital, and were excluded on a
case-by-case basis if they were:
a. Presented for a disorder that would make them unsuited to leaving wards/nursing care dur-
ing their stay in the hospital, or too painful/uncomfortable to be handled.
b. Known to be aggressive and therefore not suitable for handling
c. Isolated from the general hospital population for infection control
d. Already recruited to a separate clinical trial/study within the hospital (N.B. no other studies
investigated respiratory disease, obesity, GI disease or other breed-specific studies were
ongoing at this time that may have biased enrolment)
The owners of all remaining dogs (n = 700) were approached in the waiting room before
their consultation, to request consent.
Identification of BOAS cases
Dogs referred to the RVC SARH for BOAS were diagnosed based on clinical history, physical
examination, and examination of their upper airway under general anaesthesia. BOAS fre-
quently goes unnoticed by owners because they perceive that its signs are ‘normal’in their dog
or in brachycephalic breeds, being of early onset, highly prevalent and long-lasting [22]. There-
fore, to ensure that no dogs were erroneously classed as unaffected, dogs referred to the hospital
for unrelated reasons were also assessed for BOAS. Examination of the upper airways under
anaesthesia was not feasible for these dogs for ethical and practical reasons, including the
financial cost to the owner. Instead, therefore:
1. All study dogs were examined for stenotic nares (narrowed nostrils). This external abnor-
mality is comparatively simple to diagnose compared with the invasive diagnosis of internal
airway abnormalities. However, the severity of stenosis normally involves a subjective visual
assessment [44]. In some dogs stenosis may be mild, while in others can result in the neces-
sity to almost continually breathe with the mouth open [45]. The ‘nares ratio’was quantita-
tively calculated for all dogs, as previously published [22].
2. Questionnaires were given to all owners, with regard to their dog’s behaviour, health and
lifestyle. The frequency of respiratory difficulties and the severity of abnormal respiratory
sounds were requested in four activity scenarios: (i) at rest, e.g. while lying down awake;
(ii) while gently walking, e.g. walking around the house; (iii) during activity/exercising e.g.
on a walk, whilst playing; and (iv) while asleep. The degree of owner-reported respiratory
Impact of Facial Conformation on Canine Health: Airway Disease
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 6/21
difficulty and respiratory noise in the four scenarios was later quantified into a composite
score, the “Owner Reported Breathing”(ORB) score, out of 40 [22].
3. Clinical History: To avoid misclassification of cases, all study dogs’clinical histories including
their physical examination findings at the RVC SARH were examined to identify the presence
of other respiratory or cardiac disorders that may have contributed to their ORB score.
ORB score and nares ratio values from formally diagnosed dogs were used as a diagnostic
cut-off point to classify affected versus unaffected dogs. These thresholds were thus based on
the lowest ORB score observed in a formally affected dog (8/40) and the upper value of 95%
confidence interval of the nares ratio for the formally affected dogs (0.30; i.e. the widest nares
of any formally affected dog). As such, dogs with ORB scores >8 and nares ratio values <0.30
were classed as affected.
Ethics statement
This study was approved by the Royal Veterinary College’s Ethics and Welfare Committee
(URN 2010 1054).
Statistical analysis
Data were analysed using generalised linear mixed models for binary outcomes in R, using lmer
from the lme4 package. Being affected by BOAS was the binary response variable in all models.
Relevant morphometric predictors (all individual measures including absolute and relative
parameters e.g. absolute muzzle length and CFR), BCS, weight and age were modelled as continu-
ous fixed effects. Neck girth was included as an additional morphometric predictor because it
predicts a comparable human respiratory disorder, obstructive sleep apnoea syndrome (OSAS)
[46–50]. The presence of a nasal fold (wrinkle) was also investigated as a predictor, in case it was
an external manifestation of excess soft tissue internally. The principal component indicative of
body size (PC1) was also taken into account in case of a scaling effect on BOAS risk, as has been
found with at least one other conformational disorder [21], along with the principal component
indicative of skeletal thickness/broadness (PC2). Breed was included as a random effect, with all
cross breeds coded plainly as ‘cross breed’due to the unknown parentage of many of these dogs.
This random effect took into account the genetic non-independence of multiple members of the
same breed in the study population, and possible demographic and environmental factors. Non-
morphometric predictors i.e. signalment: age, sex, neuter status, Parker genetic groups [51]and
Kennel Club groupings [9] were tested in all models.
Multicollinearity was checked for in all models, identified from inflated standard errors in the
models, and thus avoided. Model fit was assessed using the deviance and Akaike's information
criterion. From the model output, equations were used to calculate the probability of being
affected by BOAS at different values of CFR, using breed-specific random effects to compare dif-
ferent breeds’risks. Estimates were only calculated for the CFR range exhibited in that breed, for
reasons of biological plausibility. For the variables held constant in the model whilst the fixed
effect under investigation was varied, the mean value was used for that breed, to represent an
average member of the breed, and BCS was held constant at 5, to represent an ideal weight dog.
Results–Study 1
Population demographics and clinical status
In the referral population, 700 dogs represented 97 breeds (this population has been described
in detail in Packer et al.[21]). All owners consented for their dog to be involved in the study.
Impact of Facial Conformation on Canine Health: Airway Disease
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 7/21
Most dogs were pure bred (87%), with 40% male neutered, 17% male entire, 32% female neu-
tered and 10% female entire. The mean ± SE age was 5.17 ± 0.13 years, and mean ± SE weight
(kg) was 21.5 ± 0.55, with 46% of dogs being overweight (BCS>5/9).
Of the 700 dogs recruited, 70 were categorised as affected by BOAS. Thirty five were for-
mally diagnosed, and a further 35 dogs met both of the non-invasive inclusion criteria. There
were no significant differences between formally diagnosed vs. criteria-diagnosed dogs for cra-
niofacial ratio (mean: 0.19, SE: 0.02 vs. mean: 0.15, SE: 0.01, Mann-Whitney U = 508, p<0.05),
nares ratio (mean: 0.25, SE: 0.03 vs. mean: 0.18, SE: 0.01, Mann-Whitney U = 496, p>0.05) or
breed distribution (X
2
= 20.8, p>0.05) and thus their results are combined as one ‘affected’cat-
egory due to their inherent similarities. Twelve breeds were affected (Table 2). Affected cross
breeds included two Bulldog crosses ‘Victorian Bulldogs’, and one Pug x Chihuahua ‘Chug’
and one ‘Pugalier’(Pug x Cavalier King Charles Spaniel).
Risk factors for BOAS
In study 1, as relative muzzle length increased, the proportion of dogs affected dropped steeply.
Over 80% of dogs with CFR <0.1, i.e. the muzzle less than one tenth of the cranial length, were
affected. In contrast, no dogs with CFRs of 0.5 or longer were affected; the affected dog with
the longest muzzle was a Staffordshire Bull Terrier with a CFR of 0.49. There were 236 dogs
with CFR values of 0.49 or less, falling within the conformationally ‘at-risk’range: the relatively
shortest to longest muzzle lengths affected by BOAS.
As well as shorter craniofacial ratios, increased absolute neck girths were significantly asso-
ciated with greater BOAS risk. We determined statistical significance for both factors using
generalised linear mixed models (GLMMs) for binary outcomes (including breed as a random
effect) (Table 3). No other morphometric factors were significantly associated with BOAS risk
in any models, including the presence of a nasal fold (P>0.05).
Table 2. Synopsis of breeds affected by brachycephalic obstructive airway syndrome (BOAS), morphometric risk factors and modelled BOAS
probabilities.
Breed Study 1 Study 2 Study 1 Study 2 Study 1 Study 2
n Affected
(%)
n Affected
(%)
Median
craniofacial
ratio (IQR)
Median
neck girth
(IQR)
Median
craniofacial
ratio (IQR)
Median
neck girth
(IQR)
Min-Max
predicted
BOAS risk
Min-Max
predicted
BOAS risk
Pug 32 88% 32 91% 0.08 (0.06) 32.2 (4.82) 0.12 (0.06) 31.8 (3.10) 0.69–0.97 0.48–0.95
French Bulldog 13 70% 4 75% 0.19 (0.06) 33.0 (6.25) 0.18 (0.05) 35.3 (3.70) 0.73–0.89 0.30–0.85
Bulldog 16 63% 6 33% 0.22 (0.11) 42.2 (7.58) 0.25 (0.08) 43.8 (9.75) 0.38–0.74 0.26–0.88
Boston Terrier 6 83% 2 50% 0.14 (0.04) 30.2 (2.55) 0.23 28.2 0.21–0.72 -
Japanese Chin 0 - 10 80% - - 0.04 (0.06) 23.8 (3.38) - 0.84–0.96
Pekingese 3 67% 3 0% 0.12 31.3 0.11 28.0 0.50–0.65 0.66–0.79
Dogue de
Bordeaux
6 67% 1 0% 0.36 (0.03) 55.5 (9.15) 0.36 55.1 0.22–0.47 -
Griffon
Bruxellois
2 50% 20 10% 0.13 22.6 0.15 (0.06) 24.2 (3.23) - 0.11–0.64
Boxer 13 18% 4 50% 0.31 (0.06) 41.0 (5.85) 0.30 (0.07) 38.2 (8.42) 0.24–0.55 0.02–0.31
Shih Tzu 13 8% 7 43% 0.20 (0.07) 28.5 (3.25) 0.22 (0.05) 29.2 (5.00) 0.27–0.55 0.04–0.45
Chihuahua 5 40% 3 0% 0.34 (0.17) 20.0 (1.20) 0.41 19.2 0.01–0.05 0.02–0.26
CKCS 26 4% 11 18% 0.39 (0.07) 31.2 (4.85) 0.36 (0.05) 29.2 (4.10) 0.05–0.32 0.01–0.19
Affenpinscher 1 0% 31 10% 0.20 21.1 0.23 (0.08) 23.6 (4.30) - 0.04–0.36
Staffordshire
Bull Terrier
16 6% 2 0% 0.50 (0.07) 39.2 (4.92) 0.45 (0.02) 38.8 (6.72) 0.04–0.05 -
doi:10.1371/journal.pone.0137496.t002
Impact of Facial Conformation on Canine Health: Airway Disease
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 8/21
In line with policy requests for quantitative limits to conformational traits to be available for
inclusion in Breed Standards [34,35], we used the GLMM model estimates to generate breed-
specific predictions for BOAS risk across the CFR spectrum (Fig 2A; Table A in S1 File). For
example, Pugs with the shortest observed CFR of 0.03 had a predicted BOAS risk of 0.95, com-
pared with almost half that risk, 0.48, when the CFR was 0.21, the most moderate CFR for this
breed (when neck girth was 32cm, the breed mean).
From the model predictions, the top three highest risk breeds were identified as the Pug,
French Bulldog and Bulldog. When represented graphically, clear differences emerge between
the different brachycephalic breeds with regard to BOAS risk (Fig 2A). Each curve incorporates
information on each breed’s CFR range (length of line), random effect coefficient and mean
neck girth.
Materials and Methods–Study 2
Morphometrics
Dogs were measured in an identical manner to Study 1.
Recruitment of dogs
Between July 2012 and April 2013, brachycephalic dogs (n = 154) were recruited from non-
referral populations for Study 2: breeders (79%), first opinion veterinary practice (14%), and
rescue centres (7%). Inclusion criteria for this study were individual dogs:
1. With a craniofacial ratio <0.5 (as measured by RMAP). This value represents the longest
relative muzzle length observed in an affected dog in Study 1.
2. That based on their temperament were suitable for handling for ~30 minutes and comfort-
able with a standard veterinary physical examination including rectal temperature
measurement.
Both cross and pure breed brachycephalic dogs of all ages and gender were eligible for this
study. To reflect the varied nature of the non-referral population, dogs were recruited from a
variety of sources in the South East of England. These included:
1. Breeders. Dogs currently active in the breeding (and/or showing) population were consid-
ered of high priority for this study, as they have the most genetic influence on the next gen-
eration of their respective breed. Breed clubs of those breeds considered at high-risk due to
their morphology (i.e. CFR<0.5), that were poorly represented or not included in the first
study were targeted specifically. This included the Affenpinscher, Griffon Bruxellois,
Pekingese, Japanese Chin and Tibetan Spaniel. Appropriate Breed clubs in the South East
Table 3. Mixed model results for Studies 1 and 2 demonstrating risk factors for brachycephalic obstructive airway syndrome (BOAS). Body condi-
tion score and neuter status were non-significant in Study 1, and were excluded from the final model.
Variable Study 1 Study 2
OR SE z P OR SE z P
Intercept 4.36 0.89 1.65 0.09 0.0015 2.10 -3.09 0.002
Craniofacial ratio 0.000001 2.06 -7.85 <0.001 0.0000003 3.83 -3.95 <0.001
Neck girth 1.05 0.02 2.04 0.04 1.09 0.04 2.01 0.04
Body condition score - - - -2.86 0.33 3.18 0.01
Neutered - - - - 5.66 0.70 2.48 0.01
doi:10.1371/journal.pone.0137496.t003
Impact of Facial Conformation on Canine Health: Airway Disease
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 9/21
and London regions were contacted. In addition, to increase the number of all brachyce-
phalic breeds that were currently used in breeding and showing, individual breeders in the
South East and London were identified and contacted via the Kennel Club listings of mem-
bers of the Assured Breeder Scheme.
2. Rehoming centres. This population was anticipated to incorporate several numerically large
breeds in the overall UK canid population that had previously been underrepresented in the
RVC SARH referral population. This includes the Staffordshire Bull Terrier (n = 16) and its
crosses. This breed represented only 2.35% of the referral population sample, but represents
the 2
nd
most popular breed in the UK, based on microchip data, and 6
th
based on Kennel
Club rank in 2008 [52]. Two centres; Dogs Trust Salisbury and Kenilworth were visited to
include dogs in the unowned brachycephalic population.
3. First opinion veterinary practice. This population was anticipated to incorporate brachyce-
phalic dogs with non-referral level disorders, and healthy dogs, for example those presented
for routine vaccinations. This population could potentially include dogs with mild BOAS
that were not considered in need of further investigation or treatment. A large first opinion
veterinary practice (Northolt, UK) volunteered as a study centre. Promotional material
including posters to be displayed in the waiting room and consulting rooms were produced
to advertise the study to relevant clients.
Fig 2. (a) (b) Predicted probability of brachycephalic dog breeds being affected by brachycephalic
obstructive airway syndrome (BOAS) across relevant craniofacial ratio (CFR) and neck girth ranges.
The risks across the CFR spectrum are calculated by breed using GLMM equations based on (a) Study 1
referral population data and (b) Study 2 non-referral population data. For each breed, the estimates are only
plotted within the CFR ranges observed in the study populations. Dotted lines show breeds represented by
<10 individuals. The breed mean neck girth is used for each breed (as stated in Table 2). In (b), the body
condition score (BCS) = 5 (ideal bodyweight) and neuter status = neutered.
doi:10.1371/journal.pone.0137496.g002
Impact of Facial Conformation on Canine Health: Airway Disease
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 10 / 21
Identification of BOAS cases
As diagnosis under general anaesthesia was not possible for these cases, close observation of
clinical signs before and after a gentle exercise challenge was instead used to assess the presence
and severity of BOAS. For example, the presence and intensity of ‘stertor’, a low-frequency
snoring sound, can easily be recorded and, on fluoroscopy, the sound has previously been
shown to correlate with transient inhalation of an elongated soft palate into the larynx [30].
To systematically assess BOAS clinical signs, a new standardised clinical examination proto-
col was therefore devised and carried out by the same assessor for all study dogs (MST):
1. The dog’s respiratory noise was noted at rest, and recorded for 30 seconds using an audio
recording device (Sony ICD-UX200 Digital Voice Recorder and Sony ECM-CS10 Zoom
Clip-Style Microphone), with the microphone attached to the dog’s collar, directly below
the mouth.
2. Pre-exercise, the presence and frequency of the following clinical signs and behaviours were
noted: open mouth breathing, dyspnoea, cyanotic mucous membranes, abnormal respira-
tory noise, choking and gagging episodes, yawning, postural adjustments e.g. stretched
neck, nasal discharge and sneezing.
3. The presence of internal, referred respiratory noise was detected via auscultation using a
recording stethoscope (Littmann 3200 Electronic 12 Track Stethoscope: 3200BU12) over
the trachea (caudal to the larynx) and the thorax.
4. Temperature was measured rectally, pulse and respiratory rate were recorded and capillary
refill time was assessed.
5. The dog was walked at a steady walking pace for 5 minutes. If at the RVC SARH this was
over a set route around the hospital car park. If this was at a breeder’s residence, at a first
opinion practice, or at a Dogs Trust centre, an appropriate 5 minute route was advised by
the owner or dog’s carer. The walk was curtailed if the dog appearedto show excessive signs
of respiratory distress.
6. On return from the walk, steps 1–4 were repeated in order to assess any change in respira-
tion behaviour and physiology in comparison to pre-exercise (See Video A in S3 File)
Following this assessment, the veterinary assessor (MST) then answered the following ques-
tion for each dog: In your professional opinion is this dog affected by BOAS? [YES]/[NO]
As syncope (collapse) was unlikely to be seen within this clinical examination, owners were
questioned as to whether this had occurred in their dog’s history. All owners were questioned
as to whether their dog had undergone any previous surgery to correct for BOAS, and to avoid
incorrectly attributing clinical signs of other disorders to BOAS where possible, all owners
were questioned regarding other known respiratory conditions in their dog.
As an ethical note, in dogs where a significant clinical problem was present, further investi-
gation by a veterinarian was recommended to the owner, strongly so if the problems were
severe (e.g. marked dyspnoea or history of syncope).
Statistical analysis
Statistical analyses were performed in an identical manner to Study 1.
Impact of Facial Conformation on Canine Health: Airway Disease
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 11 / 21
Results–Study 2
Population demographics and clinical status
In the brachycephalic non-referral population, 154 dogs represented 19 different breeds, with
94% pure bred. Of the 147 purebred dogs, 132 (90%) were Kennel Club registered. Female dogs
were overrepresented (65%), with the majority of both sexes being entire (66%), reflecting the
breeding dogs included in this population. The mean ± SE age was 4.12 ± 0.23 years, and
mean ± SE weight (kg) was 9.44 ± 0.79, with 57% of dogs being overweight (BCS>5/9).
In total, 59 (38.3%) were categorised as affected by BOAS, representing 10 breeds (Table 2).
Affected cross breeds included two Staffordshire Bull Terrier crosses (3%) and one Mastiff
cross (2%). No owners reported that their dogs had undergone previous surgery to correct for
BOAS.
Risk factors for BOAS
In study 2, both shorter craniofacial ratios and thicker neck girths were confirmed to increase
BOAS risk, as in study 1 (Table 3). In addition, a further two lifestyle factors were discovered
that independently predicted an increased BOAS risk: being more overweight and being neu-
tered. As before, we used GLMM model parameters to estimate breed-specific probabilities of
BOAS across the CFR spectrum (Fig 2B and Table B in S1 File), revealing marked similarities
between model estimates from Studies 1 and 2. From the model predictions, the highest risk
breeds were again identified as the Pug, French Bulldog and Bulldog.
The predicted effects of neck girth and body condition were relatively subtle compared with
CFR (Figure A in S2 File). To demonstrate the effect of neck girth, a neutered Pug with the
breed average CFR (0.11) and neck girth (32 cm) had a predicted BOAS risk of 0.93; however,
if neck girth was reduced by 5cm the predicted risk decreased slightly to 0.89. Considering
body condition in the same ‘average’Pug, if BCS increased by just 1 point to 6 (slightly over-
weight), the predicted risk rose from 0.93 to 0.98.
Study comparison
Model predictions were markedly similar between the two populations. To demonstrate this
similarity, for a dog of non-specified breed with a neck girth of 20cm, at CFR 0.1, the predicted
BOAS probability was 0.69 for Study 1, compared with 0.68 for Study 2. Likewise, for CFR 0.2,
the predicted BOAS probability for Study 1 was 0.31, and for Study 2 it was 0.32.
Conformation vs breed affiliation
Fig 3 shows that flat facial conformation, aside from other aspects of breed affiliation, plays a
major role in determining whether or not a dog has BOAS. The drop in the proportion of
affected dogs with higher CFRs was steep regardless of whether all breeds were considered (Fig
3A and 3C), or just those breeds and crossbreeds comprising affected individuals, i.e. when we
limited the genetic background solely to affected breeds (Fig 3B and 3D). The direct effect of
facial conformation is all the more evident in Fig 3, because the relationship persists despite the
fact that not all brachycephalic breeds are closely related to each other [53]. This gives little rea-
son to suppose they would share a genetic trait predisposing them to BOAS by common
descent, other than brachycephaly itself.
Discussion
Our results confirm that brachycephaly is a risk factor for BOAS and for the first time quantita-
tively demonstrate that more extreme brachycephalic conformations are at higher risk of
Impact of Facial Conformation on Canine Health: Airway Disease
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 12 / 21
BOAS than more moderate morphologies; BOAS risk increases sharply in a non-linear manner
as relative muzzle length shortens. Although susceptibility varies greatly between breeds (Fig
2A and 2B), these results suggest that breeding towards relatively longer CFRs may indeed
reduce BOAS risk within affected breeds. In addition to this finding, increased neck girth and
obesity were found to be further exacerbating factors for BOAS. Despite being studied in two
different populations, the results showed marked similarities, indicating the strong biological
nature of the relationship between morphology and disease.
This study has enabled the first quantitative estimates associating conformation with the
risk of BOAS, as requested by the Council of Europe (1995). We have demonstrated the associ-
ation across a variety of breeds, independent of breed-genetic factors, which were accounted
for through the random effect of breed in the mixed model analysis. As such, general risk factor
estimates have been modelled, that can be used to guide breeding decisions. Specific estimates
have also been modelled for a variety of high-risk brachycephalic breeds. These results can be
used to introduce quantitative limits to the degree of brachycephaly prescribed by breed stan-
dards, to encourage breeding for more moderate craniofacial morphology, in order to reduce
the prevalence and severity of BOAS. Interestingly, the three most affected breeds in both of
our studies are the same as the three most affected in a previous study on a Belgian referral
population [11]. Breeders of these breeds and others can use these estimates to see how high-
risk their current breeding stock are, and plan matings utilising this information to reduce
BOAS risk of in future offspring by aiming for safer conformations. Stakeholders can now con-
sider these values and decide what they believe is an ‘acceptable’disease-risk.
Brachycephaly as a risk factor for BOAS
As asserted by Oechtering (2010), there are ‘inevitable’major consequences of markedly reduc-
ing the bony framework of an organ [54]. Our results show that breeding for extreme craniofa-
cial morphology results has a profound effect upon the risk of obstructive upper airway disease
in domestic dogs. This scaling effect had not previously been demonstrated, with little objective
evidence that more extreme brachycephalic dogs were at increased risk [11]. The particularly
Fig 3. Percentage of dogs affected by BOAS by craniofacial ratio (CFR) category (a-d). Graphs (a) and
(c) represent all dogs in populations 1 (n = 700) and 2 (n = 154), respectively. Graphs (b) and (d) only
represent dogs of breeds and their crosses that were affected by BOAS in population 1 (n = 174) and 2
(n = 141), respectively. The breed-restricted population demonstrates the effects of conformation whilst
keeping the genetic and environmental background as uniform as possible. The marked risk of CFRs <0.20 is
clearly demonstrated in both studies, with >50% of dogs affected.
doi:10.1371/journal.pone.0137496.g003
Impact of Facial Conformation on Canine Health: Airway Disease
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 13 / 21
strong relationship between BOAS and brachycephalic phenotypes suggests that, although
BOAS susceptibility varies between breeds (Fig 2), breeding away from extreme brachycephaly
would reduce BOAS risk generally. That said, whether breeds can feasibly be selectively bred
for lower risk morphologies relies on there being sufficient existing phenotypic variability in
the breed. That is not always the case, e.g. the highest CFR in the Japanese Chin in our studies
was just 0.13, with an associated predicted BOAS probability of 0.83. Judicious out-crossing to
introduce ‘safe’conformations into a breed is controversial, but might be required on the basis
of animal welfare in populations with extremely high prevalences of BOAS and limited mor-
phological variation.
Foreshortened CFR is the key observable trait that has been selected for in the genetically
complex brachycephalic conformation [55–57], and it is likely that this external measure is a
proxy for aspects of internal morphology that skeletal foreshortening has had a pathological
impact upon. Reduced facial bone morphology without concurrent and proportional reduction
of the structures held within it is likely to have led to relative elongation and thickening of the
soft palate, and relative oversize of the turbinates. Additional internal factors may predispose
individuals or specific breeds to increased airway resistance e.g. a low glottic index [58] and
angles of the internal facial bones e.g. the hard palate. It is possible that each breed has different
internal risk factors that are incorporated within their random effect statistic in the models pre-
sented here, causing them to have differing risks of BOAS. Further study of the internal mor-
phology of the airways could investigate the relationship between these internal structures and
external muzzle foreshortening. Nevertheless, from an applied point of view, the CFR is an
important variable because dog breeders and buyers will more easily be able to select healthy
dogs on the basis of their externally visible conformation.
Despite the strong relationship between CFR and BOAS risk, it is likely that other genetic
and environmental factors will also contribute to the likelihood and severity of BOAS in indi-
vidual dogs. We found some exceptional individuals that were unaffected by BOAS despite
extreme brachycephaly (e.g. 18 individuals in Study 1, and 29 in Study 2, appeared unaffected
despite having CFRs <0.2), that may have anatomical and/or physiological adaptations.
Brachycephaly has concurrently led to a ‘space problem’for the canine brain [59], with ana-
tomical changes observed in brachycephalic dogs such as ventral pitching of the primary
longitudinal brain axis, a ventral shift in the position of the olfactory lobe [59], and more per-
pendicular development of the cranium relative to the facial axis [60] hypothesised to be adap-
tations representing a biological solution to this problem. It is possible that such anatomical
adaptations to solve the airway ‘space problem’associated with brachycephaly exist, and may
be selectable via respiratory assessment, but this requires further investigation of exceptional
individuals.
Due to ethical and financial limitations not allowing for internal airway examination of all
study dogs, BOAS was diagnosed in study 1 based on the nares ratio and ORB score. It is possi-
ble that some dogs with BOAS may have lacked stenotic nares but had other airway abnormali-
ties (e.g. an elongated soft palate) and a high ORB score, but were classified as unaffected.
Further study of BOAS populations could identify whether certain conformations are associ-
ated with specific features of BOAS or not, to ascertain the extent to which this may have
affected the results presented here. Regarding diagnosis in Study 2, it is possible that owners
may have been dishonest with their answers regarding previous surgeries to correct for BOAS,
which may have led to some dogs being erroneously classed as unaffected by BOAS. In addi-
tion, as some dogs were recruited from rescue centres, full veterinary histories may not be avail-
able. Although clinical signs can improve post-surgically, many ‘corrected’dogs are not
considered ‘normal’[45], and may still be restricted in their activities compared with unaf-
fected animals, thus may still be assessed as affected. As dogs were diagnosed without direct
Impact of Facial Conformation on Canine Health: Airway Disease
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 14 / 21
visualisation of the pharynx, it is possible that stertor was caused by other pathologies; how-
ever, as these can be assumed to be unrelated to cranial conformation, this would have reduced
our ability to detect a relationship between conformation and classification as affected by
BOAS.
Neck girth as a risk factor for BOAS
Thick neck girth and obesity have not previously been implicated in BOAS; however, both are
risk factors for OSAS in humans. Thick neck circumference predicts OSAS more effectively
than measures of obesity (e.g. BMI, waist: hip ratio) [46–50]. It is unclear to what degree neck
girth is independent of obesity in dogs, or to what extent environmental and genetic factors
contribute. If there is a strong genetic component to neck girth, selective breeding away from
thicker necks would help combat BOAS. In several high-risk Kennel Club breeds a ‘thick’neck
is explicitly encouraged [61–63]. This is additionally encouraged by the American Kennel Club
in the Pug and French Bulldog [64,65], with more extreme references to a ‘very thick’neck in
the Bulldog [66]. This may have led to artificial selection for disproportionate fat deposition in
this region. As such, breed-specific quantitative data outlining maximum neck girth values
may be a further strategy to reduce BOAS risk.
Neck girth was only a significant predictor of BOAS when included as an absolute measure,
and not when normalised against other measurements (including width of the head and chest,
chest girth and PC1). It is important to remember that the models included multiple predic-
tors, so neck girth was significant in the context of all the other predictors in the model, i.e.
once CFR, Breed and other factors were also taken into account. Thus, the models suggest that
absolute neck girth is important for a given breed with a given CFR. It is possible that the abso-
lute weight of the tissue in the neck, rather than the relative proportions of it to the airway, has
an effect on its capacity to compromise the airway. That is, the same relative neck girth might
differ in its capacity to impinge on relevant airway structures. This is also seen in human
OSAS, with an absolute neck circumference greater than 16 inches in a woman or greater than
17 inches in a man correlating with an increased risk for the disorder [46], with increasing
neck girth correlating with the severity of apnoea [48]. It should be noted that neck girth was a
relatively small effect when compared to CFR, and thus a high neck girth alone may not be suf-
ficient to cause BOAS. Further study of the necks of brachycephalic dogs, including dynamic
CT, may elucidate the role of neck girth in airway compromise [67].
Obesity as a risk factor for BOAS
Regarding obesity, generalised adipose tissue deposition could well include the palate, tongue
and tissues surrounding the airways. Indeed, obesity narrows the pharyngeal airway in Zucker
rats, and increases upper airway collapsibility [68]. Furthermore, in humans, weight loss can
alleviate OSAS [69], so maintaining a lean body condition in dogs with BOAS is potentially
important. Obesity may have further general effects on the respiratory system of the dog, with
barometric whole body plethysmography demonstrating that obese dogs had a significantly
decreased tidal volume (per kg) and significantly increased respiratory rate [70] compared with
non-obese dogs. Whether obesity is a causal factor in the development of BOAS, or an exacer-
bating factor for an existing case cannot be elucidated from our data. Indeed, it is possible that
high body condition scores in BOAS affected dogs are a consequence of BOAS limiting their
abilities to exercise and thus increasing their likelihood of obesity. A longitudinal study is
required to determine whether dogs became obese before or after exhibiting BOAS clinical
signs [71]. Obesity is prevalent in the general canine population, with an estimated 20–40% of
dogs affected [72]. In the non-referral population of brachycephalic dogs, 56.5% of dogs were
Impact of Facial Conformation on Canine Health: Airway Disease
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 15 / 21
overweight with a BCS of over 5, and 10% of dogs even had a BCS of over 7. Due to the poten-
tial exacerbating effects upon BOAS, keeping dogs at a healthy body weight is advocated.
Neutering
Neutering as a risk factor for BOAS is also a novel finding in Study 2. It is possible that this is
not a causal biological effect; rather, it could be a circular association, with BOAS causing with-
drawal from breeding and showing (thus leading to neutering) in show-dogs due to rules in
place to discourage unhealthy dogs in the showing and breeding population (e.g. the Kennel
Club initiative ‘Breed Watch’warns against breathing difficulties in several brachycephalic
breeds). As such, neutering becomes associated with BOAS in the show and breeding-dog
heavy population of Study 2. This would then lead to a disproportionate number of BOAS
affected dogs being neutered, and would be an encouraging sign of effort already being made in
the UK breeding community to avoid breeding from affected dogs. This finding may also be an
artefact of the unusual population studied here; outside of the veterinary environment, and
with a high percentage of show dogs included. If this were to be a biological effect, explanations
for this effect include neutering being a risk factor for obesity [73], which was also found to be
a risk factor for BOAS. Effects of sex hormones on the respiratory system appear an unlikely
cause. As such, from these data, keeping dogs entire is not advocated as part of the prevention
strategy for BOAS in individual dogs; however, neutering dogs affected by BOAS may be help-
ful to avoid its perpetuation in future generations. Although unlikely to be widely adopted
without policy level intervention, veterinarian Harvey [12] insisted that if dogs are to undergo
BOAS surgery, then they must also be neutered at the same time.
Reducing BOAS risk
If society wanted to eliminate BOAS from the domestic dog population entirely then based on
these data a quantitative limit of CFR no less than 0.5 (approximately describing the CFR of an
average Staffordshire Bull Terrier) would need to be imposed. However, with the current popu-
larity of brachycephalic dogs both in the UK and internationally [74], this is unlikely to be
implemented as it would require the cessation of breeding of many popular breeds. Also it
would unnecessarily exclude breeding from many moderately brachycephalic dogs that were
actually free of BOAS. Even if society were to decide that, say, a 50% risk of BOAS were accept-
able, this would mean ceasing to breed from dogs with a CFR of around 0.2 or less, approxi-
mately describing an average French Bulldog. Many breeds include few if any individuals with
CFRs above this threshold, so they would struggle to survive such a policy and the breeds
might effectively be ‘banned’. Banning is controversial, with a recent survey finding approxi-
mately as many UK stakeholders in favour of banning affected breeds immediately as those
entirely against banning [37].
If society wanted to reduce BOAS risk, but not ban any existing breeds, then an even more
moderate strategy could be adopted. Several approaches could be used towards breeding
towards more moderate, lower-risk morphologies, each of which may have strengths and
weaknesses and may be differentially supported by stakeholders involved in this issue [37]:
1. Selecting only those dogs with more moderate, lower-risk morphologies for breeding. The
further amendment of breed standards to promote lower-risk morphologies and penalise
high-risk, extreme morphologies (potentially including quantitative limits) may aid this
approach,
2. Health screening of morphologically extreme dogs to help select only those that are free of
BOAS for breeding,
Impact of Facial Conformation on Canine Health: Airway Disease
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 16 / 21
3. Developing genetic tests to highlight high and low risk animals e.g. individuals within
high risk breeds with or without elongated soft palates. Genetically testing for the
anatomical abnormalities of BOAS may not be the optimal solution as these features may
be strongly linked with skull morphology, and thus this may not be feasible strategy
and/or
4. For breeds lacking sufficient individuals with moderate morphologies, judicious out-cross-
ing to increase health and phenotypic diversity. This approach would require the necessary
cooperation from kennel clubs.
Conclusions
We present an example of how focussed selective breeding for one desired extreme morphol-
ogy can result in an unintentional pathology detrimental to animal welfare. Our results quanti-
tatively demonstrate for the first time how breeding for flatter faces dramatically increases the
risk of chronic airway obstruction in domestic dogs. The phenomenon of a desired conforma-
tion directly impacting on normal function is likely to extend to all domesticated species with
malleable phenotypes. As such, this study not only has implications for dog breeding in the
UK, but domestic animal breeding programmes internationally. Although relaxed functional
demands may have facilitated diversification of companion dogs’skull morphology [33], health
should be one of the primary considerations when breeding companion animals.
Supporting Information
S1 File. Predicted probabilities of being affected by brachycephalic obstructive airway syn-
drome (BOAS) by breed across the brachycephalic craniofacial ratio (CFR) spectrum. Data
is presented from Study 1 (referral population) data (Table A in S1 File) and Study 2 (non-
referral population) data (Table B in S1 File).
(DOCX)
S2 File. Predicted probability of a dog being affected by brachycephalic obstructive airway
syndrome (BOAS) across the brachycephalic craniofacial ratio (CFR) for three neck girths.
Data presented from Study 1 (Figure Aa in S2 File) and Study 2 (Figure Ab in S2 File).
(DOCX)
S3 File. Video of signs of brachycephalic obstructive airway syndrome (BOAS).
(DOCX)
S1 Dataset. Raw data from study 1.
(XLSX)
S2 Dataset. Raw data from study 2.
(XLSX)
Acknowledgments
We thank UFAW for the Research and Project Awards and Dogs Trust for the Canine Welfare
Grant funding this work and the RVC/BBRSC for the Quota PhD Studentship funding RMAP.
We thank the owners, breeders and dogs participating in the study for their time and co-opera-
tion, and the staff of the Clinical Investigations Centre and the Royal Veterinary College Small
Animal Referral Hospital, Dogs Trust rehoming centres and the Mandeville Veterinary Hospi-
tal for assisting with data collection. We also thank BVetMed elective students and Centre for
Impact of Facial Conformation on Canine Health: Airway Disease
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 17 / 21
Animal Welfare technicians for assistance with data collection. We are grateful to Profs Geor-
gia Mason and Alan Wilson, and Drs Richard Bomphrey, Nigel Raine, and Alex Weir for their
comments. The paper was internally approved for submission (Manuscript ID number
VCS_00631).
Author Contributions
Conceived and designed the experiments: RMAP AH CCB. Performed the experiments:
RMAP MST. Analyzed the data: RMAP CCB. Contributed reagents/materials/analysis tools:
RMAP MST. Wrote the paper: RMAP AH MST CCB. Diagnosed clinical cases: MST.
References
1. Wayne RK, editor (2001) Consequences of domestication: morphological diversity of the dog. Oxon,
UK: CABI Publishing. 43–60 p.
2. Asher L, Diesel G, Summers JF, McGreevy PD, Collins LM (2009) Inherited defects in pedigree dogs.
Part 1: Disorders related to breed standards. The Veterinary Journal 182: 402–411. doi: 10.1016/j.tvjl.
2009.08.033 PMID: 19836981
3. Summers JF, Diesel G, Asher L, McGreevy PD, Collins LM (2010) Inherited defects in pedigree dogs.
Part 2: Disorders that are not related to breed standards. The Veterinary Journal 183: 39–45. doi: 10.
1016/j.tvjl.2009.11.002 PMID: 19963415
4. Oltenacu P, Broom D (2010) The impact of genetic selection for increased milk yield on the welfare of
dairy cows. Animal Welfare 19(S): 39–49.
5. Knowles TG, Kestin SC, Haslam SM, Brown SN, Green LE, Butterworth A, et al. (2008) Leg Disorders
in Broiler Chickens: Prevalence, Risk Factors and Prevention. Plos One 3: e1545. doi: 10.1371/
journal.pone.0001545 PMID: 18253493
6. Pollinger JP, Bustamante CD, Fledel-Alon A, Schmutz S, Gray MM, Wayne RK (2005) Selective sweep
mapping of genes with large phenotypic effects. Genome Research 15: 1809–1819. PMID: 16339379
7. Stockard CR (1941) The genetic and endocrinic basis for differences in form and behavior: as eluci-
dated by studies of contrasted pure-line dog breeds and their hybrids. Philadelphia: The Wistar Insti-
tute of Anatomy and Biology.
8. Noller C, Hueber JP, Aupperle H, Seeger J, Oechtering TH, Niestrock C, et al. (2008) New aspects of
brachycephalia in dogs and cats. Basics: insights into embryology, anatomy and pathophysiology.:
Universitat Leipzig.
9. The Kennel Club (2013) Breed Registration Statistics http://www.thekennelclub.org.uk/item/1128 8th
August 2013
10. Packer RMA, Hendricks A, Burn CC (2015) Impact of Facial Conformation on Canine Health: Corneal
Ulceration. PLoS ONE 10: e0123827. doi: 10.1371/journal.pone.0123827 PMID: 25969983
11. Njikam IN, Huault M, Pirson V, Detilleux J (2009) The Influence of Phylogenic Origin on the Occurrence
of Brachycephalic Airway Obstruction Syndrome in a Large Retrospective Study. The International
Journal of Applied Research in Veterinary Medicine 7: 138–143.
12. Harvey CE (1989) Inherited and Congenital Airway Conditions. Journal of Small Animal Practice 30:
184–187.
13. Oechtering TH, Oechtering GU, Noller C (2007) Structural characteristics of the nosein brachycephalic
dog breeds analysed by computed tomography. Tierarztliche Praxis Ausgabe K Kleintiere Heimtiere
35: 177–187.
14. Leonard HC (1956) Surgical relief for stenotic nares in a dog. Journal of the American Veterinary Medi-
cal Association 128: 530. PMID: 13319181
15. Ginn JA, Kumar MSA, McKiernan BC, Powers BE (2008) Nasopharyngeal turbinates in brachycephalic
dogs and cats. Journal of the American Animal Hospital Association 44: 243–249. PMID: 18762560
16. Wykes PM (1991) Brachycephalic airway obstructive syndrome. Probl Vet Med 3: 188–197. PMID:
1802247
17. Monnet E, editor (2003) Brachycephalic airway syndrome. Philadelphia: W.B. Saunders. 808–813 p.
18. Orsher RJ, editor (1993) Brachycephalic airway disease. 2 ed. Philadelphia: Lea & Febiger. 369–370
p.
19. Reiter A, Holt D (2012) Palate. In: Tobias K, Johnston S, editors. Veterinary Surgery Small Animal Vol-
ume II. Missouri: Elsevier Saunders. pp. 1707–1717.
Impact of Facial Conformation on Canine Health: Airway Disease
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 18 / 21
20. White R (2012) Surgical management of laryngeal collapse associated with brachycephalic airway
obstruction syndrome in dogs. J Small Anim Pract 53: 44–50. doi: 10.1111/j.1748-5827.2011.01156.x
PMID: 22122300
21. Packer RMA, Hendricks A, Volk HA, Shihab NK, Burn CC (2013) How Long and Low Can You Go?
Effect of Conformation on the Risk of Thoracolumbar Intervertebral Disc Extrusion in Domestic Dogs.
PLoS ONE 8: e69650. doi: 10.1371/journal.pone.0069650 PMID: 23894518
22. Packer RMA, Hendricks A, Burn CC (2012) Do dog owners perceive the clinical signs related to confor-
mational inherited disorders as 'normal' for the breed? A potential constraint to improving canine wel-
fare. Animal Welfare 21: 81–93.
23. Hendricks JC, editor (1995) Recognition and treatment of congenital respiratory tract defects in brachy-
cephalics. Philadelphia: WB Saunders. 892–894 p.
24. Aron DN, Crowe DT (1985) Upper Airway-Obstruction—General-Principles and Selected Conditions in
the Dog and Cat. Veterinary Clinics of North America-Small Animal Practice 15: 891–917.
25. Dupre GP (2008) Brachycephalic Syndrome: New Knowledge, New Treatments. 33rd World Small Ani-
mal Veterinary Association Congress. Dublin, Ireland.
26. Farquharson J, Smith D.W. (1942) Resection of the soft palate in the dog. Journal of the AmericanVet-
erinary Medical Association 100: 427–430.
27. Schlotthauer CF (1929) Congenital defects in the stomach of a dog. J Amer Vet Med Ass 75: 370–371.
28. Roedler FS, Pohl S, Oechtering GU (2013) How does severe brachycephaly affect dog’s lives? Results
of a structured preoperative owner questionnaire. The Veterinary Journal 198: 606–610. doi: 10.1016/
j.tvjl.2013.09.009 PMID: 24176279
29. Wiestner TS, Koch DA, Nad N, Balli A, Roos M, Weilenmann R, et al. (2007) Evaluation of the repeat-
ability of rhinomanometry and its use in assessing transnasal resistance and pressure in dogs. Ameri-
can Journal of Veterinary Research 68: 178–184. PMID: 17269884
30. Hendricks JC, Kline LR, Kovalski RJ, O'Brien JA, Morrison AR, Pack AI (1987) The English bulldog: a
natural model of sleep-disordered breathing. Journal of Applied Physiology 63: 1344–1350. PMID:
3693167
31. Poncet CM, Dupre GP, Freiche VG, Estrada MM, Poubanne YA, Bouvy BM (2005) Prevalence of gas-
trointestinal tract lesions in 73 brachycephalic dogs with upper respiratory syndrome. Journal of Small
Animal Practice 46: 273–279. PMID: 15971897
32. Knecht CD (1979) Upper airway obstruction in brachycephalic dogs. Compendium on Continuing Edu-
cation for the Practicing Veterinarian 1: 25–30.
33. Drake AG, Klingenberg CP (2010) Large-Scale Diversification of Skull Shape in Domestic Dogs: Dis-
parity and Modularity. American Naturalist 175: 289–301. doi: 10.1086/650372 PMID: 20095825
34. Bateson P (2010) Independent Inquiry into Dog Breeding http://breedinginquiry.files.wordpress.com/
2010/01/final-dog-inquiry-120110.pdf 1 February 2011
35. Council of Europe (1995) Resolution on the breeding of pet animals, Multilateral Consultation of parties
to the European Convention for the protection of pet animals (ETS 123), March 1995 in Strasbourg,
Document CONS 125(95)29, Council of Europe, F 67075 Strasbourg-Cedex.
36. Farrow T, Keown AJ, Farnworth MJ (2014) An exploration of attitudes towards pedigree dogs and their
disorders as expressed by a sample of companion animal veterinarians in New Zealand. New Zealand
Veterinary Journal: 1–7.
37. Packer R, Hendricks A, Burn C (2014) Conference Report: Building Better Brachycephalics 2013.
North Mymms: Royal Veterinary College. 1–36 p.
38. Sutter NB, Mosher DS, Gray MM, Ostrander EA (2008) Morphometrics within dog breeds are highly
reproducible and dispute Rensch's rule. Mammalian Genome 19: 713–723. doi: 10.1007/s00335-008-
9153-6 PMID: 19020935
39. Regodón S, Vivo JM, Franco A, Guillén MT, Robina A (1993) Craniofacial angle in dolicho-, meso- and
brachycephalic dogs: radiological determination and application. Annals of Anatomy—Anatomischer
Anzeiger 175: 361–363.
40. Evans HE (1993) Miller's Anatomy of the Dog: Saunders.
41. Schmidt M, Neumann A, Amort K, Failing K, Kramer M (2011) Cephalometric measurements and deter-
mination of general skull type of cavalier king charles spaniels. Vet Radiol Ultrasound 4: 436–440.
42. Carrera I, Dennis R, Mellor DJ, Penderis J, Sullivan M (2009) Use of magnetic resonance imaging for
morphometric analysis of the caudal cranial fossa in Cavalier King Charles Spaniels. American Journal
of Veterinary Research 70: 340–345. doi: 10.2460/ajvr.70.3.340 PMID: 19254145
Impact of Facial Conformation on Canine Health: Airway Disease
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 19 / 21
43. Purina (2012) Assessing your dog's body condition http://www.purina.co.uk/content/your-dog/helping-
to-keep-your-dog-healthy/managing-your-dog%27s-weight/assessing-your-dog%27s-body-condition
1 November 2012
44. Brown D, Gregory S (2005) Brachycephalic airway disease; Brockman D, Holt D, editors: BSAVA.
45. Hendricks JC (1992) Brachycephalic airway syndrome. Vet Clin North Am Small Anim Pract 22: 1145–
1153. PMID: 1523786
46. Davies R, Stradling J (1990) The relationship between neck circumference, radiographic pharyngeal
anatomy, and the obstructive sleep apneoa syndrome. European Respiratory Journal 5: 509–514.
47. Davies R, Ali N, Stradling J (1992) Neck circumference and other clinical features in the diagnosis of
the obstructive sleep apnoea syndrome. Thorax 47: 101–105. PMID: 1549815
48. Katz I, Stradling J, Sljutsky AS, Zamel N, Hoffstein V (1990) Do Patients with Obstructive Sleep Apnea
Have Thick Necks? American Review of Respiratory Disease 141: 1228–1231. PMID: 2339843
49. Hoffstein V, Mateika S (1992) Differences in abdominal and neck circumferences in patients with and
without obstructive sleep apnoea. European Respiratory Journal 5: 377–381. PMID: 1563498
50. Steiropoulos P, Papanas N, Nena E, Antoniadou M, Serasli E, Papoti S, et al. (2010) Inflammatory
Markers in Middle-Aged Obese Subjects: Does Obstructive Sleep Apnea Syndrome Play a Role? Medi-
ators of Inflammation 2010.
51. Parker HG, Kim LV, Sutter NB, Carlson S, Lorentzen TD, Malek TB, et al. (2004) Genetic structure of
the purebred domestic dog. Science 304: 1160–1164. PMID: 15155949
52. Asher L, Buckland E, Phylactopoulos CI, Whiting M, Abeyesinghe S, Wathes C (2011) Estimation of
the number and demographics of companion dogs in the UK. BMC Veterinary Research 7: 74. doi: 10.
1186/1746-6148-7-74 PMID: 22112367
53. Parker HG, Kukekova AV, Akey DT, Goldstein O, Kirkness EF, Baysac KC, et al. (2007) Breed relation-
ships facilitate fine-mapping studies: A 7.8-kb deletion cosegregates with Collie eye anomaly across
multiple dog breeds. Genome Research 17: 1562–1571. PMID: 17916641
54. Oechtering GU (2010) Brachycephalic syndrome: new information on an old congenital disease. Veteri-
nary Focus: 2–9.
55. Haworth KE, Islam I, Breen M, Putt W, Makrinou E, Binns M, et al. (2001) Canine TCOF1; cloning, chro-
mosome assignment and genetic analysis in dogs with different head types. Mammalian Genome 12:
622–629. PMID: 11471057
56. Hunemeier T, Salzano FM, Bortolini MC (2009) TCOF1 T/Ser variant and brachycephaly in dogs. Ani-
mal Genetics 40: 357–358. doi: 10.1111/j.1365-2052.2008.01838.x PMID: 19220226
57. Bannasch D, Young A, Myers J, Truve K, Dickinson P, Gregg J, et al. (2010) Localization of Canine
Brachycephaly Using an Across Breed Mapping Approach. Plos One 5: e9632. doi: 10.1371/journal.
pone.0009632 PMID: 20224736
58. Caccamo R, Buracco P, La Rosa G, Cantatore M, Romussi S (2014) Glottic and skull indices in canine
brachycephalic airway obstructive syndrome. BMC Veterinary Research 10: 12. doi: 10.1186/1746-
6148-10-12 PMID: 24410902
59. Roberts T, McGreevy P, Valenzuela M (2010) Human Induced Rotation and Reorganization of the
Brain of Domestic Dogs. Plos One 5: e11946. doi: 10.1371/journal.pone.0011946 PMID: 20668685
60. Regodon S, Vivo J, Franco A (1993) Craniofacial angle in dolicho-, meso- and brachycephalic dogs:
radiological determination and application. Ann Anat 175: 361–363. PMID: 8363043
61. The Kennel Club (2012) Pug Breed Standard http://www.the-kennel-club.org.uk/services/public/breed/
standard.aspx?id=6164 1 November 2012
62. The Kennel Club (2010) Bulldog Breed Standard http://www.the-kennel-club.org.uk/services/public/
breed/standard.aspx?id=4084 1 November 2012
63. The Kennel Club (2009) French Bulldog Breed Standard http://www.the-kennel-club.org.uk/services/
public/breed/standard.aspx?id=4088 1 November 2012
64. American Kennel Club (2014) Pug Breed Standard http://www.akc.org/breeds/pug/breed_standard.
cfm 1st September 2014
65. American Kennel Club (2014) French Bulldog Breed Standard http://www.akc.org/breeds/french_
bulldog/breed_standard.cfm 1st September 2014
66. American Kennel Club (2014) Bulldog Breed Standard http://www.akc.org/breeds/bulldog/breed_
standard.cfm 1st September 2014
67. Yucel A, Unlu M, Haktanir A, Acar M, Fidan F (2005) Evaluation of the Upper Airway Cross-sectional
Area Changes in Different Degrees of Severity of Obstructive Sleep Apnea Syndrome: Cephalometric
and Dynamic CT Study. American Journal of Neuroradiology 26: 2624–2629. PMID: 16286412
Impact of Facial Conformation on Canine Health: Airway Disease
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 20 / 21
68. Magalang U, Farkas G, Najdzionek J, Nakano H (2000) Obese Zucker rats have narrower upper airway
compared to lean litter-mates. Am J Respir Crit Care Med 161:A87.
69. Smith PL, Gold AR, Meyers DA, Haponik EF, Bleecker ER (1985) Weight Loss in Mildly to Moderately
Obese Patients with Obstructive Sleep Apnea. Annals of Internal Medicine 103: 850–855. PMID:
3933396
70. Manens J, Bolognin M, Bernaerts F, Diez M, Kirschvink N, Clercx C (2012) Effects of obesity on lung
function and airway reactivity in healthy dogs. The Veterinary Journal 193: 217–221. doi: 10.1016/j.tvjl.
2011.10.013 PMID: 22099184
71. Robertson I (2003) The association of exercise, diet and other factors with owner-perceived obesity in
privately owned dogs from metropolitan Perth, WA. Prev Vet Med 58: 75–83. PMID: 12628772
72. Gossellin J, Wren J, Sunderland S (2007) Canine obesity: An overview. Journal of Veterinary Pharma-
cology and Therapeutics 30 1–10. PMID: 17567509
73. Lund E, Armstrong P, Kirk C, Klausner J (2006) Prevalence and risk factors for obesity in adult dogs
from private US veterinary practices. International Journal of Applied Research in Veterinary Medicine
4: 177–186.
74. The Kennel Club (2014) Breed Registration Statistics http://www.thekennelclub.org.uk/registration/
breed-registration-statistics 1st September 2014
75. The Kennel Club (2008) Pekingese Breed Standard http://www.the-kennel-club.org.uk/services/public/
breed/standard.aspx?id=6162 1 November 2012
76. American Kennel Club (2014) Pekingese Breed Standard http://www.akc.org/breeds/pekingese/
breed_standard.cfm 1st September 2014
77. The Kennel Club (2009) Griffon Bruxellois Breed Standard http://www.the-kennel-club.org.uk/services/
public/breed/standard.aspx?id=6154 1 November 2012
78. American Kennel Club (2014) Brussels Griffon Breed Standard http://www.akc.org/breeds/brussels_
griffon/breed_standard.cfm 1st September 2014
79. The Kennel Club (2009) Japanese Chin Breed Standard http://www.the-kennel-club.org.uk/services/
public/breed/standard.aspx?id=6156 1 November 2012
80. American Kennel Club (2014) Japanese Chin Breed Standard http://www.akc.org/breeds/japanese_
chin/breed_standard.cfm 1st September 2014
81. The Kennel Club (2009) Boston Terrier Breed Standard http://www.the-kennel-club.org.uk/services/
public/breed/standard.aspx?id=4083 1 November 2012
82. American Kennel Club (2014) Boston Terrier Breed Standard http://www.akc.org/breeds/boston_
terrier/breed_standard.cfm 1st September 2014
83. The Kennel Club (2009) Shih Tzu Breed Standard http://www.the-kennel-club.org.uk/services/public/
breed/standard.aspx?id=4102 1 November 2012
84. American Kennel Club (2014) Shih Tzu Breed Standard http://www.akc.org/breeds/shih_tzu/breed_
standard.cfm 1st September 2014
Impact of Facial Conformation on Canine Health: Airway Disease
PLOS ONE | DOI:10.1371/journal.pone.0137496 October 28, 2015 21 / 21