Access to this full-text is provided by Wiley.
Content available from Veterinary Record
This content is subject to copyright. Terms and conditions apply.
Received: 9 July 2020 Revised: 23 May 2021 Accepted: 30 May 2021
DOI: 10.1002/vetr.586
ORIGINAL RESEARCH
Effect of massage therapy on pain and quality of life in dogs:
A cross sectional study
Lisa M. Riley1,2Liam Satchell3Lisa M. Stilwell4Natalie S. Lenton5
1Centre for Animal Welfare, University of
Winchester, Winchester, UK
2Sparsholt College Hampshire, University
Centre Sparsholt, Winchester, UK
3Department of Psychology, University of
Winchester, Winchester, UK
4Meryll House, Canine Massage Guild,
Worcestershire, UK
5Meryll House, Canine Massage Therapy
Centre Ltd, Worcestershire, UK
Correspondence
L.M. Riley,Centre for Animal Welfare,Univer-
sity of Winchester,Sparkford Road, Winch-
ester SO22 4NR, UK.
Email: Lisa.Riley@winchester.ac.uk
Abstract
Background: Clinical canine massage involves muscle tissue manipulation
and fascial release techniques to rehabilitate injured soft tissues. Quantitative
efficacy data are lacking. This cross-sectional study aimed to determine how
dogs respond to canine massage therapy practiced by Canine Massage Guild
UK practitioners.
Methods: In 2018, case notes from a convenience sample of 527 dogs were
shared, with permission from owners, by a self-selected sample of 65 prac-
titioners. Changes in number and severity of issues for five pain indicators
(gait, posture, daily activity, behaviour, performance) and quality of life score,
reported by owner and practitioners, were investigated.
Results: Significant reductions in reported pain severity scores were recorded
for all pain indicators over successive treatments (p<0.001), with each treat-
ment causing further significant reduction in pain severity. Number of pain
indicators recorded over successive treatment sessions remained constant, in
keeping with a cohort presenting with degenerative disease and chronic pain.
All dogs and diagnostic variables responded similarly. Post-treatment a dog
was significantly more likely to have a ‘positive’ quality of life.
Conclusions: This cross-sectional study indicates canine massage therapy
may effectively reduce myofascial and musculoskeletal pain severity reported
by owners and practitioners associated with gait, posture, behavioural and
performance issues and reduction in daily activities. Although this is not a
double-blind trial, and there is no control group, this study suggests mas-
sage therapy may be a valid treatment for myofascial and musculoskele-
tal pain typically derived from muscular injuries, arthritis/other orthopaedic
conditions.
INTRODUCTION
Clinical Canine Massage Therapy (hereafter mas-
sage therapy or therapeutic massage) is achieved
by the application of a range of manual techniques
(e.g., effleurage, compression, friction, percussion and
stroking)1,2 to a dog’s fascia and muscle by skilled
practitioners3following veterinary consent or referral4
to address musculoskeletal injury, disorder and/or
disease.
Gait issues, muscle weakness, myalgia, localised
tremors and postural changes typically caused,
in dogs, by osteoarthritis, degenerative changes,
neuropathies, post-surgical trauma and joint
dysfunction,3overload muscles (or cause overuse of
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction
in any medium, provided the original work is properly cited and is not used for commercial purposes.
© 2021 The Authors. Veterinary Record published by John Wiley & Sons Ltd on behalf of British Veterinary Association
muscles) and initiate the formation of myofascial trig-
ger points (MTPs) restricting motor output, inhibiting
sensory input and reducing quality of life, leading
to incapacitating pain.5Myofascial pain syndrome
(MPS) results when MTPs activate muscle nociceptors
and neuroplastic changes in dorsal horn neurons lead
to hypersensitivity and allodynia.6Palpation is one
diagnostic tool for recognising MTPs,7yet palpation
is often overlooked during veterinary examinations3
and in veterinary practice generally.8Similarly, muscle
strain injury (in isolation or in conjunction with an
orthopaedic pathology) may contribute to, or cause,
acute lameness and is also under-diagnosed.9Ther-
apeutic massage, via palpation, aims to rehabilitate
injured soft tissues.7
Vet Rec. 2021;e586. wileyonlinelibrary.com/journal/vetr 1of10
https://doi.org/10.1002/vetr.586
2of10 Veter inary Rec ord
FIGURE 1 Potential therapeutic massage benefits
in dogs (after Corti, 2014)
Empirical studies on canine patients are sparse but
have found massage therapy mobilises endogenous
basal mesenchymal stem cells into circulation,10
which are associated with anti-inflammatory
responses and healing of hip osteoarthritis11 and
spinal cord injury.12 As part of rehabilitation ther-
apy, therapeutic massage enhanced postoperative
recovery following cervical disk herniation repair.13
However, isolating effects of canine massage therapy
remains a challenge, and in the human literature
only ’low strength evidence’ that massage therapy
had a therapeutic effect on pain was found.14 It
was anecdotally reported by Corti1that massage
therapy in small mammals (including dogs) caused
local changes in muscle condition and systemic
changes relating to activation or suppression of
endocrine, lymphatic and digestive systems, includ-
ing reduced pain response (Figure 1). Quantitative
research is required to understand the potential uses
and effects of massage therapy in canine veterinary
practice.
The current cross-sectional study aimed to inves-
tigate quantitative changes in pain (across multiple
indicators) reported by dog owners and practitioners
of the Canine Massage Guild UK (hereafter the Guild)
in a large sample of dogs who were referred by veteri-
narians due to soft tissue damage/injury or myofas-
cial/musculoskeletal pain. Data were extracted from
case files for over 500 dogs who received 1–3 treat-
ments using the Lenton Method of massage therapy.
It was hypothesised that pain, both number of issues
and severity, would significantly reduce post-massage
therapy, and quality of life would improve.
METHOD
This work was approved by the Ethics Committee,
University Centre Sparsholt UK (ref Riley: Canine
Massage Project). Guild practitioners only treat
with veterinary consent and after veterinary refer-
ral in compliance with the Veterinary Surgeons Act
1966 and the Veterinary Surgery (Exemptions) Order
2015.
Dogs
Data from a convenience sample of 527 dogs from
the UK, Ireland, Channel Isles and Spain were col-
lected (40 additional dogs’ data were omitted as these
records were incomplete). No inclusion criteria were
set, although practitioners were asked to provide data
on ‘a representative sample of dog types routinely seen
in practice’. While an a priori target sample size was
not calculated, the large sample of 527 dogs is suf-
ficient to detect even small, standardised effects of
r=0.09 for simple effects and f =0.07 for compar-
isons between successive massage therapy sessions.
Data on dog age, life stage, sex, neuter status, breed
(later coded into Kennel Club grouping) and role were
requested (Table 1). Seventy-one breeds were repre-
sented, although cross breeds were most numerous
(28%). Most dogs were ’pets’, ’adult’, neutered and not
on medication related to the referral diagnosis. Males
and females were equally represented.
Practitioner training, massage method and
treatment procedure
Guild practitioners train for 2 years in Clinical Canine
Massage Therapy delivered by Canine Massage Ther-
apy Centre Ltd (LANTRA accredited). Practitioners
must achieve high standards of competency in veteri-
nary anatomy and physiology, soft tissue, orthopaedic
and neurological pathologies and palpatory liter-
acy. Practitioners master 60 techniques spanning
Myofascial Release, Sports, Swedish and Deep Tissue
massage and the Lenton Method (Figure 2), a unique
three-tiered approach to treat musculoskeletal pain
in dogs. Tier 1 involves full body advanced palpation.
The practitioner will utilise their palpatory literacy to
assess a minimum of approximately 60 pairs of mus-
cles and regions of fascia to find frequently observed
muscular and fascial issues (including trigger points
and strains). Issues are summarised and classified in
Tier 2, body mapping. The bodymap is an anatomical
map of the canine muscular system used to precisely
locate muscular and fascial issues. Generalised tissues
Veteri nary Record 3of10
TABLE 1 Dog life history variables. N=527. Data presented in total counts followed by percentage of total N (n, %N). Age in months
Breed Kennel club group Role Age Life Stage Sex Neuter status On medication
Cross breed (149, 28) Mixed (149, 28) Pet (360, 68) Average (88.25) Adult (400, 76) Female (264, 50) Neutered (449, 85) No (309, 59)
Border Collie (66, 13) Pastoral (130, 25) Sporting (122, 23) Standard Deviation (42.2) Senior (106, 20) Male (263, 50) Entire (77, 15) Yes (218, 41)
Retriever Labrador (51, 10) Gundog (124, 24) Retired Sporting (21, 4) Adolescent (21, 4) Not listed (1, <1)
German Shepherd Dog (28, 5) Terrier (31, 6) Show (11, 2)
English Springer Spaniel (19, 4) Working (29, 6) Working (10, 2)
Golden Retriever (17, 3) Hound (28, 5) Retired Show (2, <1)
Greyhound (16, 3) Utility (20, 4) Retired Working (1, <1)
Unclassified Collie (15, 3) Toy (16, 3)
Cocker Spaniel (12, 2)
Jack Russell (6, 1) / Staffordshire Bull Terrier (6, 1)
are superficially warmed and prepared using tech-
niques from Swedish massage (Effleurage, Petrissage,
Vibrations, Frictions and Tapotements). As in sports
massage, these same techniques are subsequently
used to focus on distinct anatomical areas of injury.
Deep Tissue techniques then engage the tissue bar-
rier targeting deep layers of musculature. Through
discriminative touch the practitioner checks their
palpation against the map to improve accuracy and
palpatory literacy. Tiers 1 and 2 enable the practitioner
to formulate and clinically justify a bespoke treatment
plan. Tier 3 applies ‘the 7 Protocols’, advanced myofas-
cial release techniques which address myofascial pain
and dysfunction characterised by ‘Five Principles of
Pain’, pain associated with Gait, Posture, Daily Activ-
ities, Behaviour and (when relevant) Performance.
Practitioners use specific grip modulations and apply
force control through a series of hand positions and
access points to accurately engage the fascia, complet-
ing the treatment. The three tiers are followed in order
for all dogs, however tiers 1 and 2 lead to discern-
ment and clinical justification of the type, duration
and intensity of tier 3 techniques used; hence spe-
cific grips and force for example are tailored to the
dog (size, tolerance of therapy) and injury (location,
size, severity, nature [acute/chronic/degenerative]).
This creates an appropriate clinical treatment
method rather than a general massage relaxation
technique.
During the initial massage therapy session, follow-
ing veterinary consent, the practitioner consults with
the owner to review medical history of the dog and
learn the owner’s subjective concerns and observa-
tions of the dog. The practitioner will then conduct a
full dynamic gait analysis and a static postural anal-
ysis before commencing Tier 1 of the Lenton Method.
Pain is therefore assessed before (and throughout) Tier
2. Three sessions of clinical massage therapy are pro-
vided over 3–4 weeks to allow tissue yield. Pain re-
evaluation occurs before each session to gauge the
effects of the previous session.
Data collection
Data were collected between July-September 2018. An
information sheet, consent forms (for practitioners
and dog owners) with a standardised case report form
were emailed to Guild practitioners. No incentives
were offered. Sixty-five practitioners (a self-selected
sample) provided data. Practitioners could complete
the case report form for current or retrospective cases
using client case files. Information requested included
life and medical history of the dog, referral diagnosis,
cause and timescale that led to referral, list of pre-
senting issues and severity of each issue (measured
on Likert scale 1 [mild] to 5 [severe]) for five pain
indicators: Gait - defects in limb configuration while
walking, for example, reduced range of movement,
lame, limb abduction or rotation, hopping, crab-
bing; Posture - defects in configuration of body while
stationary, for example, kyphosis, idiopathic tremors,
4of10 Veter inary Rec ord
FIGURE 2 The Lenton Method of Canine Therapeutic Massage
atypical tail carriage, pronation or supination of paw;
Daily activities - changes in performance of routine
tasks that restrict ability to complete that task, for
example, reluctance/slowing on walks, difficulty get-
ting up or down stairs or in and out the car, difficulty
standing up or lying down; Behaviour - changes in
typical behavioural expression, or development of
atypical behaviour, for example, reluctance to be
petted, groomed or examined, self-mutilation, com-
pulsive behaviour, aggression, depression, anxiety,
agoraphobia, scooting (without anal gland issue)
and; Performance - specific to sporting (includ-
ing agility), working and gun dogs, any worsen-
ing of typical performance when actively working,
competing or exercising, for example, knocking
down poles or missing contact points, early onset
fatigue, retrieval issues, reluctance to work or exer-
cise. Qualitative data on quality of life were also
collected (practitioners were asked to summarise the
dog’s quality of life with a word or phrase of their
choice). Overall response to massage therapy as deter-
mined by the practitioner was listed as ’Yes’ (dog did
respond to treatment) or ’No’ (dog did not respond to
treatment).
As practitioners’ client case files are a combination
of veterinary notes provided on a consent form, practi-
tioners’ observations, and evaluations following appli-
cation of massage therapy and owner reports (e.g., an
account of how the dog copes with arthritic or other
pain climbing the stairs), data were derived from an
amalgamation of sources.
Statistical methods
Data were extracted into Microsoft Excel (2016) and
diagnosis-centric variables categorised by the primary
author, sense-checked by the third author (Table 2).
Categories were discussed and agreed upon without
the need for triangulation. Analysis was performed in
R v3.6.2 and Rstudio v1.2.5033. A zero-inflated Pois-
son random effect regression with the ‘glmmTMB’15
package for R was used to investigate changes in the
reported number of issues for the five pain indica-
tors (gait, posture, daily activity, behaviour and perfor-
mance) across successive therapy sessions. To inves-
tigate reported changes in pain severity for the five
pain indicators, linear mixed effect modelling with the
Veteri nary Record 5of10
TABLE 2 Categorisation of diagnosis-centric variables for meaningful statistical analysis
Diagnosis-centred variable Reported in data Category (definition)
Primary anatomical area
affected
e.g., ’fore’, ’carpus’, ’wrist’, ’shoulder’, ’elbow’ Forelimb(s)
(anything from forelimb toe to shoulder inclusive)
e.g., ’hind’, ’cruciate’, ’patella’, ’mid metatarsus’,
’pelvis’
Hindlimb(s)
(anything from hindlimb toe to hip and pelvis
inclusive)
e.g., ’shoulder and hip’ Forelimb(s) and Hindlimb(s)
(any combination of forelimb and hindlimb)
e.g., ’hernia’, ’stomach’ Ventral aspect
(Abdomen and ventral thorax)
e.g., ’disc(s)’, ’back’, ’epaxial muscles’ Musculature associated with the vertebral column
(anything associated with thoracic to lumbar
spine)
e.g., ’general trauma’, ’all over body tension’ Generalised areas
(systemic disease or no specific area of clinical
concern)
Cause of referral condition e.g., ’collision’, ’racing injury’, ’sudden’, ’spinal stroke’
’operation’
Acute trauma
(sudden onset, associated with a clear accident or
incident)
e.g., ’age-related’, ’present for a long time’,
’intermittent’
Chronic
(long term stable condition, not progressive but
could be recurring)
e.g., ’osteoarthritis’, ’elbow dysplasia’, ’hip dysplasia’,
’general slowing down’, ’worse over time’
Degenerative
(progressive condition, associated with one or
more joints)
‘lme4’16 was used and where omnibus effects were
required, ANOVA were fitted using base R.
Severity of pain indicators were conceptually and
statistically (in number and pattern) similar for all
indicators, hence for efficient analysis severity of pain
was formatted into a single factor (’common pain
score’), which excluded severity of performance as this
was not reported in most cases (Table 3), an arte-
fact of the specialist population. Common pain score
was the empirical product of the covariances between
severity of gait, posture, activity and behaviour pain
issues. The ‘psych’17 package was used to conduct
an oblimin exploratory factor analysis with a one-
factor solution, an efficient summary of the variance
in the severity measures (RMSEA =0.13, 90% CI [0.10,
0.15], TLI =0.93) and weighted by the variance in Gait
(0.84), Posture (0.73), Activity (0.76) and Behaviour
(0.46) pain severity. Common pain score is a standard-
ised score with M =0.00, SD =0.92, Min =−1.63
and Max =2.30, where a higher score reports gen-
erally higher pain severity. Dog-centric (sex, age and
if the dog were on medication) and diagnosis-centric
variable (original cause of referral, primary anatomi-
cal area effected) effects on common pain score were
analysed using a series of linear mixed models, testing
for an interaction effect with the linear model of suc-
cessive treatments.
Quality of life was qualitatively analysed using
deductive thematic analysis to achieve three themes:
Positive; Fair; Negative according to words tradition-
ally associated with indicators of poor, ambiguous, or
good welfare. Themes were converted into an ordinal
scale (1 =negative, 2 =fair, 3 =good) to allow quan-
titative analysis. Change in quality of life score over
successive treatments was evaluated using cumulative
linked mixed modelling using the ‘ordinal’.18 Due to
the number of tests being conducted, a conservative
p≤0.001 criterium was used.
RESULTS
Of the 527 dogs whose data were included in this
study, practitioners provided data on overall success
of treatment for 515 dogs. Of these, 492 dogs (95.5%)
were considered by practitioners to have ’responded
to treatment’. No discernible difference in life history
variables existed between dogs who did respond to
treatment and dogs that did not. Both groups were of
similar age (No [did not respond]: M =88.5 months,
SD =42.1 months; Yes [did respond]: M =88.4 months,
SD =42.1 months), were similar breeds (all breeds that
did not respond were represented in the list of breeds
that did respond), had proportionately similar num-
bers on medication (No 35%, Yes 42%) with similar
diagnoses (primary anatomical area affected: 1) mus-
culature of the vertebral column No 17%, Yes 15%;
Hindlimb(s) No 48%, Yes 46%; Forelimb(s) No 17%, Yes
24%; Fore and Hindlimb(s) No 9%, Yes 2%; General
Areas No 9%, Yes 12%).
For each indicator of pain, number of issues
reported over successive treatments remained con-
sistent throughout the study (Gait =2, Posture =1,
Daily Activity =2, Behaviour =1, Performance =<1)
(Table 3). Gait and daily activity issues were reported
to be more numerous than posture, behaviour and
performance issues, irrespective of treatment. In tests
of both an overall linear effect of treatment, and of
6of10 Veter inary Rec ord
TABLE 3 Descriptive statistics for the number and severity of pain issues over successive treatments. Baseline =initial presentation prior to treatment
Median (minimum/maximum) number of issues Mean (SD) severity of pain
Pain type Baseline Post-treatment 1 Post-treatment 2 Post-treatment 3 Baseline Post-treatment 1 Post-treatment 2 Post-treatment 3
Gait 2 (0, 9) 2 (0, 9) 2 (0, 9) 2 (0, 9) 2.96 (1.36) 2.53 (1.31) 2.03 (1.18) 1.61 (1.06)
Posture 1 (0, 7) 1 (0, 7) 1 (0, 7) 1 (0, 7) 2.42 (1.58) 2.14 (1.46) 1.73 (1.29) 1.45 (1.19)
Activity 2 (0, 7) 2 (0, 6) 2 (0, 6) 2 (0, 6) 2.62 (1.6) 2.16 (1.47) 1.7 (1.25) 1.40 (1.09)
Behaviour 1 (0, 5) 1 (0, 6) 1 (0, 6) 1 (0, 6) 1.83 (1.75) 1.54 (1.54) 1.15 (1.26) 0.93 (1.04)
Performance 0 (0, 7) 0 (0, 5) 0 (0, 5) 0 (0, 5) 0.90 (1.57) 0.67 (1.33) 0.51 (1.11) 0.37 (0.88)
Commonpainscore 0.46 (0.95) 0.15 (0.9) -0.21(0.79) -0.48 (0.71)
changes from treatment to treatment, no significant
differences in the number of reported issues were
found (Table 4).
For each indicator of pain, reported severity score
reduced indicating a marked improvement after treat-
ment (Table 3). After referral, the sample’s combined
median pain severity was 2.4 (moderate), after three
treatments this reduced to 1.4 (mild). Gait was con-
sistently the highest pain severity score followed by
either Daily Activity (up to the second treatment) or
Posture (after the second treatment). A highly signif-
icant decrease in reported severity of all pain indi-
cators across treatments was observed (Table 5). By
the time treatment was complete, and severity scores
for all pain indicators had approximately halved (Gait
54% reduction, Posture 60% reduction, Daily Activity
54% reduction, Behaviour 51% reduction and Perfor-
mance 41% reduction). Further, significant decreases
in reported pain severity scores were observed from
initial referral to post-treatment 1 and continued to
decline significantly with each subsequent treatment
(Table 5). Over 93% of dogs showed improvement in
reported pain severity score for one or more pain indi-
cator after one or more treatments.
Using a common pain score of overall severity, a
main effect of medication (est. =−0.32, p<0.001)
was found. An interaction effect between treatments
and medication (est. =−0.83, p<0.001) high-
lights reported pain severity for dogs on medication
(M =0.86, SD =0.86) was most different to reported
pain severity for dogs not on medication (M =0.17,
SD =0.90, t(479.95) =8.82, p<0.001) during initial
assessment. It is reasonable to expect dogs in more
pain would be on medication. Once treatment was
complete, this difference was smaller though still sig-
nificant (dogs on medication M =−0.24, SD =0.72;
dogs not on medication M =−0.66, SD =0.64,
t(375.55) =6.30, p<0.001).
Increased dog age generally predicted increased
general pain severity measured by common pain score
(est. =0.01, p<0.001), although there was no inter-
action effect over successive treatments (est. =−0.00,
p=0.086), suggesting older dogs presented with
more severe pain on initial assessment but both
older and younger dogs responded to treatment sim-
ilarly as both groups experience decreased pain sever-
ity. Sex had neither a general effect on common
pain score (est. =0.12, p=0.134) nor were there
any significant effects of sex on common pain score
over successive treatments (est. =−0.03, p=0.029),
both male and female dogs respond equally well to
treatment.
Dogs with degenerative causes of referral had sig-
nificantly higher common pain scores (M =0.33,
SD =0.90) than those with chronic (M =0.06,
SD =0.90; est. =0.63, p<0.001) and acute (M =0.00,
SD =0.93; est. =0.32, p=0.003) causes. There was no
significant difference between the acute and chronic
groups (est =0.01, p =0.903). The original cause did
not interact with the improvement in common pain
score across successive treatments (all est ≤0.0.03,
Veteri nary Record 7of10
TABLE 4 Overall linear and categorical treatments zero-inflated poisson random effects models predicting number of pain issues over successive treatments across the pain indicators. Estimate (p
value)
Treatments as categories
Treatments as scaleaVs baselinebVs post-treatment 1bVs post-treatment 2b
No. issues Overall Post-treatment 1 Post-treatment 2 Post-treatment 3 Post-treatment 2 Post-treatment 3 Post-treatment 3
Gait -0.02 (0.265) -0.01 (0.727) -0.03 (0.490) -0.05 (0.288) -0.02 (0.731) -0.03 (0.466) -0.02 (0.691)
Posture -0.01 (0.356) -0.01 (0.842) -0.03 (0.583) -0.04 (0.423) -0.02 (0.727) -0.03 (0.541) -0.01 (0.782)
Activity -0.01 (0.389) -0.03 (0.536) -0.04 (0.397) -0.04 (0.421) -0.01 (0.817) -0.01 (0.833) 0.00 (0.990)
behaviour -0.01 (0.665) -0.02 (0.775) -0.02 (0.702) -0.03 (0.679) -0.01 (0.922) -0.01 (0.888) -0.00 (0.963)
Performance -0.06 (0.080) -0.15 (0.134) -0.17 (0.092) -0.17 (0.096) -0.02 (0.841) -0.03 (0.806) -0.01 (0.957)
aTreatment points treated as a scale of 1–4 for a linear omnibus test.
bTreatment points treated as discrete categories (baseline [initial presentation prior to treatment], post-treatment 1, post-treatment 2 and post-treatment 3) for change over entire treatment period.
TABLE 5 Overall linear and categorical treatments liner mixed models predicting severity of pain issues over successive treatments across pain indicators.Estimate(pvalue)
Time as categories
Treatment as Scale aVs baselinebVs post-treatment 1bVs post-treatment 2b
Pain severity Overall Post-treatment 1 Post-treatment 2 Post-treatment 3 Post-treatment 2 Post-treatment 3 Post-treatment 3
Gait -0.50 (***) -0.44 (***) -1.00 (***) -1.48 (***) -0.56 (***) -1.04 (***) -0.48 (***)
Posture -0.37 (***) -0.29 (***) -0.74 (***) -1.07 (***) -0.45 (***) -0.78 (***) -0.33 (***)
Activity -0.47 (***) -0.46 (***) -0.98 (***) -1.38 (***) -0.52 (***) -0.92 (***) -0.40 (***)
behaviour -0.33 (***) -0.30 (***) -0.70 (***) -0.98 (***) -0.41 (***) -0.68 (***) -0.27 (***)
Performance -0.16 (***) -0.20 (***) -0.35 (***) -0.49 (***) -0.15 (***) -0.29 (***) -0.14 (***)
***p<0.001.
aTreatments treated as a scale of 1–4 for a linear omnibus test.
bTreatment points treated as discrete categories (baseline [initial presentation prior to treatment], post-treatment 1, post-treatment 2 and post-treatment 3) for change over entire treatment period.
8of10 Veter inary Rec ord
FIGURE 3 Dendrogram showing thematic analysis of
quality of life. Themes shown in larger font, followed by examples
of words recorded by practitioners
all p≥0.152). While some variability in common
pain score was noted between primary affected areas
(F(5,1926) =9.70, p=0.001), the primary area affected
did not affect common pain score over successive
treatments (F(15, 1926) =0.55, p=0.915).
Quality of life ratings were coded into predeter-
mined themes of ’Positive’, ’Fair’, ’Negative’ (Figure 3).
With quality of life as a linear predictor, the omnibus
effect of successive treatments was a significant, posi-
tive, predictor (est. =0.58, p<0.001), suggesting a gen-
eral improvement in practitioner reported quality of
life over the course of massage therapy. A significant
improvement in reported quality of life between ini-
tial assessment and treatment 1 (est. =0.67, p<0.001),
between first and second treatments (est. =0.54,
p<0.001) and between second and third treatments
(est. =0.48, p<0.001) was found. At initial assessment,
40% of dogs were considered to have a positive quality
of life, this increased to 66% post first treatment, 83%
post second treatment and 92% post third treatment.
DISCUSSION
This study, in line with the aims, has shown massage
therapy may be an effective treatment for myofas-
cial/musculoskeletal pain in dogs, as reported by
practitioners, and owners. All types of dog and
anatomical areas affected responded to massage
therapy as reported severity of pain indicators signif-
icantly reduced in all cases, as hypothesised. Efficacy
of massage therapy was quick – reported severity
scores reduced after the first massage treatment and
all subsequent treatments. Reported severity scores
reduced across all indicators of pain.
Several veterinary practitioners and academics have
anecdotally reported a reduction of pain following
massage therapy in dogs,1,2,3 and massage therapy is
a recognised part of rehabilitation therapy19 where, it
is speculated, massage can reduce pain by decreas-
ing mechanoreceptor sensitivity or reducing inflam-
mation (by increasing fluid removal from tissues into
lymph and circulatory systems).20 Clinical trials of
the combined therapeutic effects of acupuncture and
massage therapy to alleviate canine musculoskele-
tal pain showed a highly significant improvement in
pain measured by owner reports of daily activity.21 In
this published trial, owners were blinded to the treat-
ment protocol. The authors used a robust experimen-
tal design, but causal relationship between massage
therapy (alone) and pain reduction could not be iso-
lated. In equines, clinical experiments showed one
treatment of massage increased mechanical nocicep-
tive thresholds by 8%.22 Similarly, a meta-analysis of
clinical trials involving humans with musculoskeletal
disorders23 found massage therapy reduced shoulder
and knee pain. The results achieved currently corrob-
orate previous research.
Of 527 dogs, 23 were considered by practitioners
to have shown no discernible response to massage
therapy. There were no identifiable differences in dog
type between those who responded to canine mas-
sage and those who did not respond. For example,
both groups had a sample that sustained new injury
(unrelated to the massage therapy), both responders
and non-responders were diagnosed by the referring
veterinarian to have degenerative and therefore pro-
gressive causes of soft-tissue damage. Many factors
affect pain, for example, MPS (and therefore hyper-
sensitivity to pain) can occur,6and owner compliance
with home-care plans issued by Guild practitioners
can differ. Recording these and similar factors that
could account for differences in overall response to
massage therapy were beyond the scope of the cur-
rent study but warrant further research. It may be
that this type of massage therapy did not suit these
dogs – this is unlikely due to the bespoke nature of
the massage application as tier 3 techniques are tai-
lored to each dog, but it cannot be ruled out. These
dogs may have had undiagnosed conditions. Muscu-
loskeletal therapies are aimed at rehabilitation and
support of muscular, skeletal and neurological condi-
tions. How effective the therapy is can often depend on
which system the problem is weighted, if the problem
is muscular better results will typically be seen than if
the problem is neurological due to the nature of the
disorder.
Regardless of species, studies on the therapeutic
effects of massage have investigated relative changes
in pain indicators rather than absolute changes. If the
aim of massage therapy is to reduce pain this is appro-
priate, however it seems reasonable to consider that
some indicators of pain could be resolved if the fas-
cial tissues containing MTPs are fully mobilised. To the
authors’ knowledge, the current study is the first to
consider if the number of pain-related issues reduces
Veteri nary Record 9of10
after massage therapy. Across the five pain indicators,
no significant reduction in reported number of pain
issues was found. This could be a genuine therapeutic
outcome. Massage therapy is used to treat over-loaded
muscles,3not the cause(s) of muscle overload. Mul-
tiple disorders lead to muscular and fascial dysfunc-
tion, either by direct injury or as secondary responses
to chronic degenerative joint issues like osteoarthri-
tis. Treatment of the soft tissues cannot directly treat
underlying bone disease but aims to address overcom-
pensation and habitual patterns of tension. In this
study, a cause for referral was included in extracted
data for 379 dogs. Of these, 250 dogs (66%) had
chronic or degenerative disease/issues. The lack of a
significant difference in pain issues may alternatively
be due to an artefact of data recording. Severity in
several cases was recorded as 1 (mild) indicating the
pain issue remained relevant, yet additional notes pro-
vided by the practitioner indicated the issue was no
longer relevant. The case report form did not allow
severity to be graded ‘zero’. It was intended that if
an issue were no longer relevant, practitioners would
leave the form blank, although this was not specified
and evidently caused confusion. If the issue was listed
by the practitioner as 1 or more, it was considered an
issue and included in analysis, regardless of supple-
mentary information from the practitioner.
Independency of data may be questioned by some
in the current study as Guild practitioners provided
data on their own applications of canine massage
from client files and extracted the assessments of
owners. Clinical reports are an important and well-
established data source in veterinary research,24 and
in the veterinary profession, owner reports are an
integral part of diagnostics. Epstein reports, for dogs,
owner assessment is scientifically and clinically the
most useful chronic pain scoring tool.25 The owners’
role in canine massage therapy is therefore important
as symptoms that lead to referral are often the observ-
able outcomes of pain, and both veterinary and mas-
sage practitioner assessments rely heavily on owner
reports of changes in a dog’s movement, behaviour
and daily activities.26 Pain, particularly chronic pain,
is a genuine welfare concern, diminishing quality
of life, mobility and inhibiting performance of daily
activities;25 hence the range of pain indicators used
currently and need for massage practitioners to be
skilled and knowledgeable.3,19 While pain scores in
this study were reported by massage practitioner and
owner, evidence of bias is limited. It is reasonable
to assume both parties would want to see significant
improvement in pain severity scores following treat-
ment, as was recorded here. It is further reasonable to
consider both parties would want to see a significant
reduction in number of pain-related issues across the
five pain indicators (if an improvement was clinically
possible given the referral diagnosis), yet this was not
found. Also, number of cases, of practitioners and the
diversity of dogs/referral conditions makes subcon-
scious bias unlikely. The large sample size and multi-
ple dog types studied suggest these results are gener-
alisable beyond the sample.
Current results suggest canine massage using the
method described may be a valid treatment for
myofascial/musculoskeletal pain; however here the
causal effects of massage are not discernible – without
experimental control groups or conditions, the causes
of significant effects remain attributable to massage
therapy, time in recovery or an artefact of increased
human contact which is important for dog well-
being.27 Time is an unlikely contributing factor given
the number of dogs with degenerative conditions
associated with secondary fascial tissue formation,
thus time is likely to worsen any presenting issues
or pain severity. In this study, dogs with degenera-
tive issues consistently scored higher for pain severity
compared to dogs with acute or chronic issues. Neither
owners nor practitioners were blind to the treatment
protocol; Lane and Hill acknowledge how difficult this
wouldbetoachieve,
21 however, as dogs on medica-
tion in the current study consistently presented with
higher pain scores and showed greater reductions in
pain scores after massage therapy, this would cor-
roborate the need for a double-blind clinical trial to
further understand the efficacies of canine massage
therapy.
Musculoskeletal issues are the third most common
reason (after enteropathic and dermatological issues)
to attend a veterinary clinic for mid-level diagnoses.28
Yet muscle pain remains poorly recognised in vet-
erinary medicine.8Previous research2and this cur-
rent study show the versatility and potential range of
effectiveness of canine massage therapy means this is
becoming an increasingly relevant treatment option
for dogs exhibiting myofascial/musculoskeletal pain.
CONCLUSIONS
Across multiple dog types and diagnoses, clini-
cal canine massage therapy was associated with
a significant reduction in practitioner, and owner
reported pain severity scores in all pain indicators
(Gait, Posture, Behaviour, Daily Activity, Performance).
Although the number of pain-related issues reported
was consistent throughout treatment, practitioners
reported significantly more dogs had a positive qual-
ity of life post treatment. Double blind clinical trials
are needed to elucidate cause and effect; however, this
large-scale quantitative analysis suggests canine ther-
apeutic massage may be an efficacious treatment for
muscle pain in dogs.
ACKNOWLEDGEMENT
We sincerely thank the practitioners who provided
data for this study – we are grateful for your commit-
ment and patience. We also thank the dog owners who
gave permission for their dog’s data to be included in
this study. This study did not receive any financial sup-
port.
10 of 10 Veteri nary Record
CONFLICT OF INTEREST
The authors declare that there is no conflict of interest
that could be perceived as prejudicing the impartiality
of the research reported.
Two members of The Canine Massage Guild (L.M.
Stilwell and N.S. Lenton) co-authored this study. N.S.
Lenton is the Director of the Canine Massage Ther-
apy Centre, Founder and Chairperson of the Canine
Massage Guild and developed the Lenton Method.
L.M. Stilwell provided technical support to L.M. Riley
when diagnoses variables were being categorised. Nei-
ther L.M. Stilwell or N.S. Lenton was involved with
analysing data or interpretation of results.
AUTHOR CONTRIBUTIONS
Riley led the research (including data extraction) and
wrote the manuscript. Satchell performed data analy-
sis in Rstudio and helped to write the Results section.
Stilwell and Lenton helped to develop the Introduc-
tion and liaised with Guild members to provide data.
Stilwell sense-checked categorisation of diagnostic-
centric variables and both Stilwell and Lenton pro-
vided information on the Lenton Method and massage
protocols. All authors agree contributions as stated
here.
ORCID
Lisa M. Riley https://orcid.org/0000-0003-1918-
4623
REFERENCES
1. Corti L. Massage therapy for dogs and cats. Top Companion
Anim Med. 2014;29(2):54–7.
2. Formenton MR, Pereira MAA, Fantoni DT. Small animal mas-
sage therapy: a brief review and relevant observations. Top
Companion Anim Med. 2017;32(4):139–45.
3. Wall R. Introduction to Myofascial Trigger Points in Dogs. Top
Companion Anim Med. 2014;29(2):43–8.
4. Royal College of Veterinary Surgeons. Code of professional con-
duct for veterinary surgeons supporting guidance. 2020. www.
rcvs.org.uk/. Accessed 13 May 2020.
5. Simons DG, Travell JG, Simons LS. Travell & Simons’ myofascial
pain and dysfunction: upper half of body (Vol. 1). 2nd rev edn.
Baltimore: Lippincott Williams & Wilkins; 1999.
6. Gerwin RD, Dommerholt J, Shah J. An expansion of Simons’
integrated hypothesis of trigger point formation. Curr Pain
Headache Rep. 2004;8:468–75.
7. Shah JP, Thaker N, Heimur J. Myofascial trigger points
then and now: a historical and scientific perspective. PM R.
2015;7(7):746–61.
8. Lindley S, Watson P. BSAVA manual of canine and feline reha-
bilitation, supportive and palliative care: case studies in patient
management. Quedgeley, Gloucester: British Small Animal Vet-
erinary Association; 2010.
9. Nielsen C, Pluhar GE. Diagnosis and treatment of hind limb
muscle strain injuries in 22 dogs. Vet Comp Orthop Traumatol.
2005;18(4):247–53.
10. Loghmani MT, Metzger T, Kim MS. Electro-acupuncture and
massage mobilize mesenchymal stem cell into circulation in
canines. J Bodyw Mov Ther. 2018;22(4):871.
11. Cuervo B, Rubio M, Sopena J, Dominguez JM, Vilar J, Morales
M et al. Hip osteoarthritis in dogs: a randomized study using
mesenchymal stem cells from adipose tissue and plasma rich
in growth factors. Int J Mol Sci. 2014;15(8):13437–60.
12. Park JH, Kim DY, Sung IY Choi GH, Jeon MH, Kim KK, et al.
Long-term results of spinal cord injury therapy using mes-
enchymal stem cells derived from bone marrow in humans.
Neurosurgery. 2012;70(5):1238–47.
13. Jeong IS, Rahman MM, Choi GC, Seo BS, Lee GJ, Kim S, et al.
A retrospective study of canine cervical disk herniation and
the beneficial effects of rehabilitation therapy after ventral slot
decompression. Vet Med (Praha). 2019;64(6):251–9.
14. Miake-Lye IM, Mak S, Lee J, Tuger T, Taylor SL, Shanman R,
et al. Massage for pain: an evidence map. J Altern Complement
Med. 2019;25(5):475–502.
15. Brooks ME, Kristensen K, van Benthem KJ, Magnusson A, Berg
CW, Nielsen A, et al. glmmTMB balances speed and flexibil-
ity among packages for zero-inflated generalized linear mixed
modelling. R J. 2017;9(2):378–400.
16. Bates D, Mächler M, Bolker B, Walker SC. Fitting linear mixed-
effects models using Ime4. J Stat Softw. 2015;67(1):1–48.
17. Revelle W. Psych: procedures for psychological, psychometric,
and personality research. Evanston, Illinois: Northwestern Uni-
versity; 2019.
18. Christensen RHB. Ordinal—regression models for ordinal data.
R package version 2019.12.10. https://CRAN.R-project.org/
package=ordinal. Accessed 1st May 2020.
19. Hesbach AL. Manual therapy in veterinary rehabilitation. Top
Companion Anim Med. 2014;29(1):20–3.
20. Millis D, Levine D. Canine rehabilitation and physical therapy.
2nd rev edn. St. Louis: Saunders Elsevier; 2014.
21. Lane DM, Hill SA. Effectiveness of combined acupuncture and
manual therapy relative to no treatment for canine muscu-
loskeletal pain. Can Vet J. 2016;57(4):407–14.
22. Sullivan K, Hill A, Haussler KK. Effect of different treatment
modalities on spinal nociceptive thresholds in horses. In: Pro-
ceedings of the 53rd Annual Convention of the American Asso-
ciation of Equine Practitioners. 2007:58–59.
23. Bervoets DC, Luijsterburg PA, Alessie JJ, Buijs MJ, Verhagen AP.
Massage therapy has short-term benefits for people with com-
mon musculoskeletal disorders compared to no treatment: a
systematic review. J Physiother. 2015;61(3):106–16.
24. Jones-Diette JS, Brennan ML, Cobb M, Doit H, Dean RS. A
method for extracting electronic patient record data from prac-
tice management software systems used in veterinary practice.
BMC Vet Res. 2016;12(1):239.
25. Epstein ME. Assessing chronic pain in dogs. Today Vet Pract.
2013;3:32–5.
26. The Canine Massage Guild. Vet information. 2020. www.k9-
massageguild.co.uk/. Accessed 2 May 2020.
27. Menor-Campos DJ, Molleda-Carbonell JM, López-Rodríguez R.
Effects of exercise and human contact on animal welfare in a
dog shelter. Vet Rec. 2011;169(15):388–91.
28. O′Neill DG, Church DB, McGreevy PD, Thomson PC, Brod-
belt DC. Prevalence of disorders recorded in dogs attend-
ing primary-care veterinary practices in England. PLoS One.
2014;9(3):1–16.
How to cite this article: Riley LM, Satchell L,
Stilwell LM, Lenton NS. Effect of massage
therapy on pain and quality of life in dogs: A
cross sectional study. Vet Rec. 2021;e586.
https://doi.org/10.1002/vetr.586
Available via license: CC BY-NC 4.0
Content may be subject to copyright.
