Content uploaded by Johannes P A M van Loon
Author content
All content in this area was uploaded by Johannes P A M van Loon on Nov 17, 2017
Content may be subject to copyright.
Monitoring acute equine visceral pain with the Equine Utrecht
University Scale for Composite Pain Assessment (EQUUS-COMPASS)
and the Equine Utrecht University Scale for Facial Assessment of Pain
(EQUUS-FAP): A scale-construction study
Johannes P.A.M. van Loon a,*, Machteld C. Van Dierendonck a,b,c
aDepartment of Equine Sciences, Faculty of Veterinary Medicine, Utrecht University, Yalelaan 114, NL-3584 CM Utrecht, The Netherlands
bVeterinary Faculty Ghent University, Salisburylaan 133, 9820 Merelbeke, Ghent, Belgium
cEquus Research and Therapy, Tolnegenweg 39, 3776PT Stroe, The Netherlands
ARTICLE INFO
Article history:
Accepted 20 August 2015
Keywords:
Colic
Equine
Pain scale
Facial expression
EQUUS-COMPASS
EQUUS-FAP
ABSTRACT
Although recognition of equine pain has been studied extensively over the past decades there is still need
for improvement in objective identification of pain in horses with acute colic. This study describes scale
construction and clinical applicability of the Equine Utrecht University Scale for Composite Pain Assess-
ment (EQUUS-COMPASS) and the Equine Utrecht University Scale for Facial Assessment of Pain (EQUUS-
FAP) in horses with acute colic. A cohort follow-up study was performed using 50 adult horses (n=25
with acute colic, n=25 controls). Composite pain scores were assessed by direct observations, Visual Analog
Scale (VAS) scores were assessed from video clips. Colic patients were assessed at arrival, and on the
first and second mornings after arrival. Both the EQUUS-COMPASS and EQUUS-FAP scores showed high
inter-observer reliability (ICC =0.98 for EQUUS-COMPASS, ICC =0.93 for EQUUS-FAP, P<0.001), while a
moderate inter-observer reliability for the VAS scores was found (ICC =0.63, P<0.001). The cut-off value
for differentiation between healthy and colic horses for the EQUUS-COMPASS was 5, and for differenti-
ation between conservatively treated and surgically treated or euthanased patients it was 11. For the EQUUS-
FAP, cut-off values were 4 and 6, respectively.
Internal sensitivity and specificity were good for both EQUUS-COMPASS (sensitivity 95.8%, specific-
ity 84.0%) and EQUUS-FAP (sensitivity 87.5%, specificity 88.0%). The use of the EQUUS-COMPASS and EQUUS-
FAP enabled repeated and objective scoring of pain in horses with acute colic. A follow-up study with
new patients and control animals will be performed to further validate the constructed scales that are
described in this study.
© 2015 Elsevier Ltd. All rights reserved.
Introduction
Animal pain and its recognition and management have re-
ceived abundant attention over the past decades. Various studies
describing tools for objective assessment of pain in farm animals
(Prunier et al., 2013), companion animals (Hellyer, 2002; Hansen,
2003) and horses (Wagner, 2010) have laid the basis for the devel-
opment of improved methods for objective pain assessment, which
in turn have contributed positively to animal welfare (Valverde and
Gunkel, 2005). Apart from species differences, pain expression is also
dependent on the type and origin of pain. Somatic pain and vis-
ceral pain, for instance, are different phenomena that manifest
differently and need to be treated differently (Robertson, 2002). Colic
is one of the most important and often diagnosed diseases in the
horse and the availability of purpose built, specific and validated
pain assessment tools would help identify colic related pain and
therefore add significantly to equine welfare and support quality
of patient care.
Simple one-dimensional pain scales like the VAS (Visual Analog
Scale) are very commonly used, but are deemed suboptimal in-
struments for pain evaluation in animals, partly because of poor
inter-observer reliability (Lindegaard et al., 2010). Composite pain
scales (CPS) offer the advantage that combining various indicators
increases sensitivity and specificity of pain assessment (Abbott et al.,
1995; Dobromylskyj et al., 2000; Prunier et al., 2013). In humans,
multidimensional pain scales have been developed for recurrent ab-
dominal pain in children (Malaty et al., 2005) and assessment of
infant pain (Cong et al., 2013). Such pain scales have been de-
scribed for different types of pain in horses like acute orthopaedic
pain (Bussières et al., 2008) and pain after colic surgery (Pritchett
* Corresponding author. Tel.: +31 30 253 1111.
E-mail address: j.p.a.m.vanloon@uu.nl (J.P.A.M. van Loon).
http://dx.doi.org/10.1016/j.tvjl.2015.08.023
1090-0233/© 2015 Elsevier Ltd. All rights reserved.
The Veterinary Journal 206 (2015) 356–364
Contents lists available at ScienceDirect
The Veterinary Journal
journal homepage: www.elsevier.com/locate/tvjl
et al., 2003; Graubner et al., 2011; van Loon et al., 2014). Recently,
a behaviour-based pain scale for horses with acute colic has been
described and validated (Sutton et al., 2013a, 2013b).
Another strategy to assess objectively the amount of pain in an
individual is to quantify facial expression. This technique has been
studied in humans (Ahola Kohut et al., 2012) and rodents (Sotocinal
et al., 2011) and has led to the development of grimace scales for
the latter species. Recently, the Horse Grimace Scale (HGS) follow-
ing surgical castration has been described (Dalla Costa et al., 2014)
and the Equine Pain Face has been described after induced exper-
imental pain (Gleerup et al., 2014).
In previous studies, the CPS as described by Bussières et al. (2008)
was used to assess pain in a cross-section of equine patients in a
referral centre (van Loon et al., 2010) and after emergency lapa-
rotomy and treatment in an intensive care unit (van Loon et al., 2014).
This pain scale, although originally developed for orthopaedic pain,
contains various elements that can also be applied to visceral pain
(van Loon et al., 2014). In the current study, the CPS was used as
the basis for the development of the Equine University Utrecht Scale
for Composite Pain Assessment (EQUUS-COMPASS), a scale aiming
at the optimal assessment of acute colic pain. Furthermore, a com-
posite facial expression pain scale was constructed, the Equine
University Utrecht Scale for Facial Assessment of Pain (EQUUS-
FAP), based on several facial expression characteristics.
The aims of the current study were (1) to assess inter-observer
variability of the EQUUS-COMPASS and the EQUUS-FAP; (2) to de-
termine cut-off points for determination between healthy and painful
animals; and (3) to assess the clinical applicability for the identi-
fication and follow-up of pain in horses with acute colic. The
hypotheses were that EQUUS-COMPASS and EQUUS-FAP would have
better inter-observer reliability than a VAS scale, would be easily
clinically applicable and would be able to differentiate between
control horses and colic patients.
Materials and methods
Animals
The study design was approved by the institutional Ethics Committee on the
Care and Use of Experimental Animals in compliance with Dutch legislation on
animal experimentation. Because the procedures used in this study only contain
behavioural observations and physiological assessments (heart rate, breathing
rate, borborygmi, rectal temperature) taken from clinical patients and therefore
are not likely to cause pain, suffering or distress or lasting harm equivalent to, or
higher than, that caused by the introduction of a needle (article 1.5f EU directive
2010/63/EU), ethical approval was obtained without an official approval number.
Furthermore, owner’s consent was obtained for all horses and ponies participating
in this study.
Twenty-five horses that had been admitted to the equine referral centre with
acute colic were included (Table 1). Twenty-five control horses (healthy mares that
were used as recipients for embryo transfer and horses that came in for regular
shoeing) that were admitted in the same period were included as well (Table 1).
All control horses were free from lameness and/or teeth problems.
The total study population consisted of 30 mares and 20 geldings. Stallions,
foals and mares with foals were excluded from the study because of possibly
disturbing effects of sexual arousal or mare–foal interaction on the assessment of
pain scores. Breeds included Warmbloods (37), Thoroughbreds (crossbreeds) (3),
Friesians (5), Irish Cob (1), Fjorden horse (1), Haflinger (1), Icelandic (1) and
Quarterhorse (1). Analgesic treatment and clinical decision-making were at the
discretion of the attending veterinarian and independent of pain scores. The
observers were not involved with day-to-day patient care and were unaware of
any analgesic treatment. All colic patients were given non-steroidal anti-
inflammatory drugs (NSAIDs) before referral to the university clinic. If α2-agonists
were required after arrival at the clinic, the horses were excluded from the study
because of possible interference with pain scores.
Equine Utrecht University Scale for Composite Pain Assessment (EQUUS-COMPASS)
The EQUUS-COMPASS is based on the CPS described by Bussières et al. (2008).
For the development of EQUUS-COMPASS, the CPS was modified by deleting pa-
rameters that are not possible to assess in horses with acute abdominal pain (e.g.
appetite) and by adding parameters that are thought to be more specific for viscer-
al pain (such as tail flicking, laying down and sounds produced as an expression of
pain like teeth grinding or moaning). The EQUUS-COMPASS is a multifactorial simple
descriptive scale (SDS) based on 14 parameters. It includes physiological param-
eters, responses to stimuli, and spontaneous behavioural parameters (Table 2). Each
of the 14 parameters can be scored from 0 to 3, leading to a total pain score range
from 0 (no signs of pain) to 42 (maximal pain score).
Equine Utrecht University Scale for Facial Assessment of Pain (EQUUS-FAP)
The EQUUS-FAP is a multifactorial SDS based on nine parameters, describing dif-
ferent elements of facial expression, like appearance of eyelids, nostrils and muscle
tone (Table 3). Each of the nine parameters can be scored from 0 to 2, leading to a
total pain score ranging from 0 (no signs of pain) to 18 (maximal pain score).
Experimental design
Observations were performed by four observers (veterinary students) who per-
formed their observations pairwise and simultaneously. The observers did not discuss
their findings. Prior to commencement of the study, all observers were given the
chance to familiarize and train themselves with the parameters in the EQUUS-
COMPASS and EQUUS-FAP using pain-free horses (not included in the study). The
observers were not blinded for the clinical diagnosis.
Patients were evaluated shortly after admission to the university hospital (T0)
when horse owners were registered to the patient database, the first morning after
admission (T12-24) and the second morning after admission (T36-48). Each obser-
vation period lasted 10 min. Scoring was performed with the animals in the colic
box, where a video camera recorded the box during scoring. These videos were used
to obtain Visual Analog Scores (VAS) on a continuous scale between 0 and 10 (Hawker
et al., 2011), performed by two observers (equine veterinarians not involved in the
treatment and blinded for time and treatment). The control horses were observed
in the same colic box and EQUUS-COMPASS and EQUUS-FAP scores were obtained
once, with simultaneous video recordings for VAS scores.
Data processing and statistical analysis
All data are expressed as medians and quartiles. Inter-observer reliability was
assessed using Intraclass Correlation Coefficients (ICC). Bland–Altman plots were
used to visually evaluate correlations and determine limits of agreement (average
difference ±1.96 standard deviation of the difference) (Bland and Altman, 1986;
Myles, 2007). Differences in scores between control animals and colic patients and
between conservatively treated animals (CT) and surgically treated or euthanized
animals (STE) were analyzed using the Mann–Whitney Utest. Cut-off values for
EQUUS-COMPASS, EQUUS-FAP and VAS were determined to obtain maximal differ-
entiation between colic patients and healthy animals and between CT and STE
treatments. Internal sensitivity, specificity, and positive and negative predictive
values were determined for EQUUS-COMPASS, EQUUS-FAP and VAS scores using
these cut-off values. Sensitivity and specificity for individual parameters of both
scores were also determined. Based on these values, weighting factors for future
validation for the different individual parameters were determined retrospectively.
When sensitivity or specificity was ≤25%, a weighting factor of 0 was applied;
between 25% and 50% the weighting factor was 1; between 50% and 75% the
weighting factor was 2; and when both sensitivity and specificity were ≥75% a
weighting factor of 3 was applied. The effects over time for both EQUUS-COMPASS
and EQUUS-FAP scores in colic patients were assessed by means of the Friedman
test. Statistical analysis was performed using SPSS version 20.0 (IBM). Statistical
significance was accepted at P<0.05.
Results
Inter-observer reliability
Fig. 1 shows the results of correlation analysis between the dif-
ferent pain scores of two independent observers. Both the EQUUS-
COMPASS and EQUUS-FAP scores showed strong and significant
correlation (ICC =0.98, P<0.001 for EQUUS-COMPASS scores,
Table 1
Data of horses that were included in the study (n=50).
Colic Control
Number of horses 25 25
Conservative treatment (CT) 15 –
Surgical treatment/euthanasia (STE) 5/5 –
Warmblood/Thoroughbred 17 23
Other breeds 8 2
Mean (±SD) weight (kg) 540 (71.8) 593 (37.6)
Mean (±SD) age (years) 11 (6.4) 9 (4.3)
357J.P.A.M. van Loon, M.C. Van Dierendonck/The Veterinary Journal 206 (2015) 356–364
ICC =0.93, P<0.001 for EQUUS-FAP scores), while a moderate but
significant correlation for the VAS scores was found (ICC =0.63,
P<0.001).
Differences between subgroups of control and colic patients
Both EQUUS-COMPASS and EQUUS-FAP scores showed signifi-
cant differences between control and colic patients (P<0.001). The
EQUUS-COMPASS showed a statistically significant difference
between CT and STE patients (P<0.01), while the EQUUS-FAP scores
between these two groups were not significantly different (P=0.84).
VAS scores were not different between control and colic patients
(P=0.051), nor between CT and STE patients (P=0.77) (Fig. 2).
Relation of breed, age and sex with pain scores
There was no relationship between EQUUS-COMPASS and EQUUS-
FAP scores and breed or sex (Figs. 3A–D). EQUUS-COMPASS scores
in the group of young horses (0–5 years) were significantly lower
than in elderly horses (>15 years) (P<0.05, Fig. 3E). EQUUS-FAP
scores between these two age groups were not significantly differ-
ent. However, the distribution of CT and STE patients was also
Table 2
Score sheet of the Equine Utrecht University Scale for Composite Pain Assessment (EQUUS-COMPASS).
Data Categories Score
Physiological data
Heart rate
24–44 beats/min
45–52 beats/min
53–60 beats/min
>60 beats/min
0
1
2
3
Respiratory rate 8–13 breaths/min
14–16 breaths/min
17–18 breaths/min
>18 breaths/min
0
1
2
3
Rectal temperature 36.9 °C–38.5 °C
36.4 °C–36.9 °C or 38.5 °C–39.0 °C
35.9 °C–36.4 °C or 39.0 °C–39.5 °C
35.4 °C–35.9 °C or 39.5 °C–40.0 °C
0
1
2
3
Digestive sounds Normal motility
Decreased motility
No motility
Hypermotility or steel band
0
1
2
3
Behaviour
Posture
Quietly standing and/or one hind leg resting, explores environment
Slightly tucked up abdomen, still explores environment (with possible unrest)
Extremely tucked up abdomen, hunched back and/or stretching of body/limbs
Does not stand or for short amounts of time (<1 min), sits on hindquarters
0
1
2
3
Laying down, rolling Does not lie down or rests lying down
Lies down in normal posture, rolls or tries to roll (once or twice/5 min)
Alternates lying down and standing, rolls or tries to roll (more than twice/5 min)
Constantly lies in an abnormal position: on its side with stretched
limbs, on its back, or does not stop rolling
0
1
2
3
Sweating No signs of sweating
Warm or damp to touch, no sweat or wet spots visible
Wet spots visible, no droplets or streams
Excessive sweating, may include streams or droplets
0
1
2
3
Tail flicking (excluding
flicking to
chase off insects)
No tail flicking
Occasional tail flicking (once or twice/5 min) and/or holds tail away from body
Frequent tail flicking (three to four times/5 min), may hold tail away from body
Excessive tail flicking (more than five times/5 min)
0
1
2
3
Kicking at abdomen Quietly standing, no kicking
Occasional kicking at abdomen (once or twice/5 min)
Frequent kicking at abdomen (three to four times/5 min)
Excessive kicking at abdomen (more than five times/5 min)
0
1
2
3
Pawing at floor (number of
episodes)
Quietly standing, does not paw at floor
Occasional pawing at floor (once or twice times/5 min)
Frequent pawing at floor (three to four times/5 min)
Excessive pawing at floor (more than five times/5 min)
0
1
2
3
Head movements No fast movements, head mostly at same height/in same direction
Occasional head movements laterally/vertically, looking at flank (once or twice/5 min)
Frequent/fast head movements laterally/vertically, looking at flank (three to four times/5 min)
Excessive head movements, excessive looking at flank (more than five times/5 min), biting at flank (more than once/5 min)
0
1
2
3
Pain sounds No audible signs of pain
Occasional teeth grinding or moaning (once or twice/5 min)
Frequent teeth grinding or moaning (three to four times/5 min)
Excessive teeth grinding or moaning (more than five times/5 min)
0
1
2
3
Overall appearance, reaction
to observer(s)
Quiet but alert, approaches/turns to observer
Alert, no reluctance to move, obvious reaction to sounds and/or movements
Restless, constantly moving, exaggerated reaction to sounds and/or movements
Stupor: the horse is not moving, head is lowered, reluctance to move
0
1
2
3
Reaction to palpation of
painful
area in the flank
No reaction to palpation
Mild reaction to palpation
Resistance to palpation
Violent reaction to (attempt to) palpation
0
1
2
3
Tota l . . ./42
358 J.P.A.M. van Loon, M.C. Van Dierendonck/The Veterinary Journal 206 (2015) 356–364
unequal in those two age groups (0–5 years: eight horses CT and
one horse STE, >15 years: three horses CT and five horses STE).
Effects over time in horses admitted with acute colic
Fig. 4 shows both EQUUS-COMPASS and EQUUS-FAP scores over
time of CT patients. For both pain scores, a significant decrease over
time was found (P<0.05).
Cut-off values between different categories of horses
Cut-off values for the EQUUS-COMPASS were 5 for differentia-
tion between controls and colic patients and 11 as cut-off value for
the distinction between CT and STE patients. Cut-off values for the
EQUUS-FAP were 4 for differentiation between controls and colic
patients and 6 for the distinction between CT and STE patients. Cut-
off values for the VAS were 2 for differentiation between controls
and colic patients and 5 as cut-off value for the distinction between
CT and STE patients. Cut-off values for individual parameters of both
EQUUS-COMPASS and EQUUS-FAP were 0 for controls and >0for
colic patients; for the differentiation between CT and STE patients
the cut-off value was 0 (one respectively >1).
Internal sensitivity and specificity of composite pain scores and their
individual parameters
Table 4 shows internal sensitivity and specificity of EQUUS-
COMPASS, EQUUS-FAP and VAS scores using the cut-off values as
described above. Sensitivity and specificity of the individual pa-
rameters and the determined weighting factors are shown in Table 5.
Discussion
This study describes the construction of both the EQUUS-
COMPASS and EQUUS-FAP, a composite pain scale and a facial
expression pain scale, respectively, for an objectified evaluation of
the severity of pain symptoms in horses with acute colic. Scale de-
velopment was supported by the high inter-observer reliability of
both scores and the clinically acceptable limits of agreement that
were found using Bland–Altman analysis. Good internal specific-
ity and sensitivity for differentiating between pain-free healthy horses
and colic patients and between colic patients that were treated con-
servatively and colic patients that were treated surgically or were
euthanased show that the scoring systems were able to differen-
tiate different pain levels (no pain, mild to moderate pain, and severe
pain). In a follow-up study, a new cohort of horses will be used to
further assess clinical applicability and to assess external sensitiv-
ity and specificity.
In the original CPS by Bussières et al. (2008) the key specific and
most sensitive behavioural indices for acute orthopaedic pain were
response to palpation of the painful area, posture and, to a lesser
extent, pawing on the floor, kicking at the abdomen and head move-
ments. For visceral postoperative pain (van Loon et al., 2014), pawing
on the floor, overall appearance, head movements and interactive
behaviour were the most important parameters of the CPS. The most
sensitive parameters of the EQUUS-COMPASS for acute colic were
the character of borborygmi, posture, sweating and reaction to ob-
server and palpation of the painful flank. The importance of
borborygmi for the assessment of severity of colic corresponds with
the results of White et al. (2005), who found significantly in-
creased odds ratios for need of surgery in animals with decreased
or no intestinal sounds.
In the current study, sensitivity and specificity were deter-
mined for all individual parameters of both the EQUUS-COMPASS
and EQUUS-FAP and these were used to determine weighting factors.
By using these weighting factors for individual parameters, sensi-
tivity, specificity and positive and negative predictive values can
potentially be improved with a limited number of parameters. Cut-
off points between healthy individuals and those suffering from colic
pain and between different levels of pain were determined by op-
timizing sensitivity and specificity for differentiation between them.
The cut-off values and weighting factors we determined in this scale
construction study will be used in a follow-up study for determi-
nation of external sensitivity and specificity.
The EQUUS-FAP could potentially be used by horse owners and
non-veterinarians after proper training, while the EQUUS-COMPASS
contains several physiological parameters (borborygmi, heart and
breathing frequency) that can be assessed only by veterinarians and
Table 3
Score sheet of the Equine Utrecht University Scale for Facial Assessment of Pain (EQUUS-FAP).
Data Categories Score
Head Normal head movement/interested in environment
Less movement
No movement
0
1
2
Eyelids Opened, sclera can be seen in case of eye/head movement
More opened eyes or tightening of eyelids. An edge of the sclera can be seen 50% of the time
Obviously more opened eyes or obvious tightening of eyelids. Sclera can be seen >50% of the time
0
1
2
Focus Focussed on environment
Less focussed on environment
Not focussed on environment
0
1
2
Nostrils Relaxed
A bit more opened
Obviously more opened, nostril flaring and possibly audible breathing
0
1
2
Corners mouth/lips Relaxed
Lifted slightly
Obviously lifted
0
1
2
Muscle tone head No fasciculations
Mild fasciculations
Obvious fasciculations
0
1
2
Flehming and/or yawning Not seen
Seen
0
2
Teeth grinding and/or moaning Not heard
Heard
0
2
Ears Position: Orientation towards sound/clear response with both ears or ear closest to source
Delayed/reduced response to sounds
Position: backwards/no response to sounds
0
1
2
Tota l . . ./18
359J.P.A.M. van Loon, M.C. Van Dierendonck/The Veterinary Journal 206 (2015) 356–364
well-educated owners. A condensed form of the EQUUS-COMPASS
without physiological parameters could also be of potential use for
non-veterinarians. Objectified multi-dimensional cut-off values for
differentiation between healthy and sick animals can help deter-
mine the necessity for veterinary help or for referral to a veterinary
clinic or, after treatment, evaluation of disease/pain progression.
Pain scores will never replace clinical decision making (with phys-
ical examination, blood results, ultrasound etc.), but can aid follow-
up of a patient and objectify responses to treatment. The importance
of pain as a prognosticating factor for the severity of colic has also
been described by others (Grulke et al., 2001; White et al., 2005),
who found strongly increased odds ratios for the need for surgery
in horses with constant pain or with pain that returned after
treatment.
Recently, in the studies by Sutton et al. (2013a, 2013b), another
concept has been applied to construct and validate pain scales for
horses that experience acute colic (Equine Acute Abdominal Pain
Scale: EAAPS-1 and EAAPS-2). These authors used a clinimetric ap-
proach to generate, select and appoint weights to items that were
involved in pain expression in horses with acute colic. Expert opinion
was used in a Delphi process to independently judge several items
that best reflected expression of pain. Although this procedure is
scientifically sound, item selection and the application of weight-
ing factors still depend on expert opinion, which can be subjective.
Furthermore, the readout parameters of these two pain scales were
based on expression of only one behavioural parameter and in-
creasing pain scores were attributed to different types of behaviour.
In EAAPS-2, the frequency of expression of the specific behavioural
items influenced the pain score. The disadvantage of the EAAPS is
that it does not take into account that factors such as breed, sex or
individual effects may influence expression of pain through differ-
ences in behavioural patterns. Composite pain scales offer the
advantage of these differences in individual expression.
In the follow-up article by Sutton et al. (2013b), both behaviour-
based pain scales (EAAPS-1 and -2) were validated by means of
digital film clips showing horses with various degrees of acute colic
and control horses. Inter-observer reliability was very good for both
EAAPS scales (ICC >0.75), while the Numerical Rating Scale (NRS)
that was used in this study for comparison showed only fair inter-
observer reliability (ICC =0.67). This is in accordance with our results,
but we found higher inter-observer reliability for the EQUUS-
COMPASS and the EQUUS-FAP (ICC >0.93) and only moderate inter-
observer reliability for the Visual Analog Scale (VAS) (ICC =0.63).
In contrast to our study, Sutton et al. (2013b) also assessed intra-
observer reliability. However, only one video was shown twice to
the observers for assessment of intra-observer reliability, which limits
the value of this analysis. With respect to sensitivity and specific-
ity for discrimination between colic cases and controls and between
Fig. 1. Scatter plot of EQUUS-COMPASS scores (A) (n=72, r=0.94, ICC =0.98, P<0.001), EQUUS-FAP scores (C) (n=72, r=0.84, ICC =0.93, P<0.001) and VAS scores (E) (n=57,
r=0.34, ICC =0.63, P<0.05), assessed by two different observers at the same moment. Bland–Altman plots of EQUUS-COMPASS (B), EQUUS-FAP (D) and VAS scores (F) show
limits of agreement of −2.8 to +2.6 for EQUUS-COMPASS, −2.2 to +2.8 for EQUUS-FAP and −2.9 to +3.4 for VAS.
360 J.P.A.M. van Loon, M.C. Van Dierendonck/The Veterinary Journal 206 (2015) 356–364
colic patients with different treatments (surgical versus conserva-
tive), the EAAPS-1 pain scale obtained best results with a sensitivity
of 89% and a specificity of 75% for discrimination between colic cases
and controls and of 70% and 57%, respectively, for discrimination
between animals with different treatments.
By means of the EQUUS-FAP, the amount of pain expressed by
horses with acute colic was assessed by means of facial expres-
sion only. Facial expression has been proven suitable for
quantification of pain in rodents (Sotocinal et al., 2011) and infants
(Ahola Kohut et al., 2012). Validation of grimace scales has been suc-
cessful for rodents and such a scale has been proven to be suitable
for horses after surgical castration in a recent study (Dalla Costa et al.,
2014). We chose to build a multifactorial numerical rating scale, with
descriptions of various aspects of facial expression such as ear and
head movements and appearance of eyelids and nostrils, and pain-
related sounds like teeth grinding, moaning and yawning. Our results
showed that facial expression was very suitable for recognition and
quantification of the severity of acute visceral pain in horses. Apart
from acute visceral pain, pilot studies have shown that the EQUUS-
FAP is also very suitable for recognition of acute dental and ocular
pain in horses (unpublished data).
One of the limitations of our study is that observers were not
blinded for clinical diagnosis. However, this is inevitable with live
observations, and because observations were strictly structured and
the observers were not involved with clinical decision-making and
treatment, this was deemed unimportant. Another limitation is that
a ‘gold standard’ for scoring acute visceral pain needed for valida-
tion of our pain scores is lacking. Therefore, we decided to obtain
video clips from all horses, taken when composite pain scores were
assessed. These videos were used to obtain VAS scores for com-
parison with the composite pain scores. However, in accordance with
other studies (Lindegaard et al., 2010; Sutton et al., 2013a), inter-
observer reliability of VAS scores was low with wide limits of
agreement and 95% confidence intervals and VAS scores proved not
useful for clinical validation of the composite pain scores.
Another possibility for better clinical validation of the pain scales
could be to assess the effect of visceral analgesics like NSAIDs or
opioids (more specifically κ-agonists) on development of pain scores
in patients with acute colic. In our study, this was not possible
because most horses with colic were referred to the university hos-
pital after being treated with an NSAID. A follow-up study carried
out in horses with acute colic at the first visit of the attending
Fig. 2. Median EQUUS-COMPASS, EQUUS-FAP and VAS scores for colic (n=25) versus control (n=25) horses (A +C+E) and surgically treated/euthanased (STE) (n=10) versus
conservatively treated (CT) (n=15) colic patients (B +D+F). Lines in boxes show median scores; boxes show 25–75th percentiles; error bars show 5–95th percentiles.
** =P<0.01, *** =P<0.001.
361J.P.A.M. van Loon, M.C. Van Dierendonck/The Veterinary Journal 206 (2015) 356–364
veterinarian would be suitable in order to test this. External vali-
dation with a new dataset of colic patients and healthy control horses
will be performed to further determine the reliability and clinical
usefulness of both pain scales and to complete their validation.
Assessment of the pain scoring systems by other veterinarians
or veterinary technicians would elucidate more on the overall added
value of the systems. Another step forward for practical implemen-
tation of the pain scoring systems will be to determine optimal
Fig. 3. Median EQUUS-COMPASS and EQUUS-FAP values for Warmbloods/Thoroughbreds (n=17) and Cold bloods (n=8: Friesians, n=3; Irish Cob, n=1; Quarterhorse, n=1;
Fjorden, n=1; Haflinger, n=1; Icelandic, n=1) animals (A +B), horses with different sexes (mares, n=13; geldings, n=12) (C +D) and horses with different ages (0–5 years,
n=9; 6–15 years, n=8; >15 years, n=8) (E +F). Lines in boxes show median scores; boxes show 25–75th percentiles; error bars show 5–95th percentiles. * =P<0.05.
Fig. 4. Median EQUUS-COMPASS (A) and EQUUS-FAP (B) values at admission to clinic (T0), at the first morning after admission (T12-24) and at the second morning after
admission (T36-48) (n=13). Lines in boxes show median scores; boxes show 25–75th percentiles; error bars show 5–95th percentiles. Friedman tests show a significant
decrease over time for both EQUUS-COMPASS (P=0.021) and for EQUUS-FAP scores (P=0.012).
362 J.P.A.M. van Loon, M.C. Van Dierendonck/The Veterinary Journal 206 (2015) 356–364
reduction of the number of parameters in both the EQUUS-COMPASS
and EQUUS-FAP while optimizing sensitivity and specificity with
the most condensed composite pain scales.
Conclusions
The use of the EQUUS-COMPASS and the EQUUS-FAP improved
objectivity in pain assessment of horses with acute colic. Because of
the high inter-observer reliability over a range of severity of pain scores,
these pain scales enabled comparisons between different observers
and so offered the possibility of monitoring the pain status of a patient
more objectively over a period of time. This is a major asset in larger
veterinary hospitals where often several attending clinicians are in-
volved in the treatment of a single patient and will also offer possibilities
for non-veterinarians to participate in evaluations.
Conflict of interest statement
This study was partly funded by Boehringer Ingelheim BV, but
this played no role in the study design, in the collection, analysis
and interpretation of data, or in the manuscript writing or submis-
sion for publication. None of the authors has any other financial or
personal relationships that could inappropriately influence or bias
the content of the paper.
Acknowledgements
We would like to thank Marianne Beikes, Larissa Vermin, Anne-
Marie Sas and Marlijn van Gent for their help with data collection
and René van Weeren for critical reading of the manuscript.
References
Abbott, F.V., Franklin, K.B.J., Westbrook, R.F., 1995. The formalin test: Scoring properties
of the first and second phases of the pain response in rats. Pain 60, 91–102.
Ahola Kohut, S., Pillai Riddell, R., Flora, D.B., Oster, H., 2012. A longitudinal analysis
of the development of infant facial expressions in response to acute pain:
Immediate and regulatory expressions. Pain 153, 2458–2465.
Bland, J.M., Altman, D.G., 1986. Statistical methods for assessing agreement between
two methods of clinical measurement. Lancet 8, 307–310.
Table 4
Internal sensitivity, specificity, positive and negative predictive value of the Equine Utrecht University Scale for Composite Pain Assessment (EQUUS-COMPASS), the Equine
Utrecht University Scale for Facial Assessment of Pain (EQUUS-FAP) and the Visual Analog Scale (VAS).
Sensitivity Specificity Positive predictive value Negative predictive value
Colic (n=25) versus controls (n=25)
EQUUS-COMPASS 95.8 84.0 85.2 95.5
EQUUS-FAP 87.5 88.0 87.5 88.0
VAS 62.5 84.0 44.0 87.5
CT (n=15) versus STE (n=10)
EQUUS-COMPASS 80.0 85.7 80.0 85.7
EQUUS-FAP 30.0 64.3 37.5 56.3
VAS 25.0 75.0 50.0 50.0
Weighting factors of individual parameters can be found in Table 5.
CT, conservatively treated animals; STE, surgically treated or euthanased animals.
Table 5
Sensitivity, specificity and weighting factors for individual parameters of the Equine Utrecht University Scale for Composite Pain Assessment (EQUUS-COMPASS) and the
Equine Utrecht University Scale for Facial Assessment of Pain (EQUUS-FAP).
Colic ↔control Weighting factoraCT ↔STE Weighting factora
Sensitivity Specificity Sensitivity Specificity
(%) (%)
EQUUS-COMPASS
Heart frequency 46 54 1 100 40 1
Breathing frequency 42 54 1 86 67 2
Rectal temperature 08 100 0 –
Borborygmi 92 96 3 80 36 1
Posture 88 72 2 44 75 1
Laying down, rolling 29 100 1 60 50 2
Sweating 54 100 2 38 100 1
Tail flicking 58 08 0 –
Kicking at abdomen 29 100 1 40 50 1
Pawing at floor 63 48 1 86 38 1
Head movements 17 80 0 –
Pain sounds 38 72 1 0 50 0
Reaction to observer 58 80 2 100 44 1
Reaction to palpation 54 80 2 50 83 2
EQUUS-FAP
Head 75 96 3 14 73 0
Eyelids 58 52 2 14 100 0
Focus 58 100 2 40 60 1
Nostrils 58 40 1 50 83 2
Lips 71 64 2 50 91 2
Muscle tone 29 92 1 0 80 0
Flehmen 0 96 0 – –
Teeth grinding 29 100 1 67 0 0
Ears 71 100 2 50 73 2
CT, conservatively treated animals; STE, surgically treated or euthanased animals.
aWhere sensitivity or specificity is ≤25%, the weighting factor =0; 25–50%, weighting factor =1; 50%–75%, weighting factor =2; ≥75%, weighting factor =3.
363J.P.A.M. van Loon, M.C. Van Dierendonck/The Veterinary Journal 206 (2015) 356–364
Bussières, G., Jacques, C., Lainay, O., Beauchamp, G., Leblond, A., Cadoré, J.L.,
Desmaizières, L.M., Cuvelliez, S.G., Troncy, E., 2008. Development of a composite
orthopaedic pain scale in horses. Research in Veterinary Science 85, 294–306.
Cong, X., McGrath, J.M., Cusson, R.M., Zhang, D., 2013. Pain assessment and
measurement in neonates. Advances in Neonatal Care 13, 379–395.
Dalla Costa, E., Minero, M., Lebelt, D., Stucke, D., Canali, E., Leach, M.C., 2014.
Development of the Horse Grimace Scale (HGS) as a pain assessment tool in
horses undergoing routine castration. PLoS ONE 9, e92281.
Dobromylskyj, P., Flecknell, P.A., Lascelles, B.D., Livingston, A., Taylor, P.,
Waterman-Pearson, A., 2000. Pain assessment. In: Flecknell, P.A.,
Waterman-Pearson, A. (Eds.), Pain Management in Animals. W.B. Saunders,
Philadelphia, PA, USA, pp. 53–79.
Gleerup, K.B., Forkman, B., Lindegaard, C., Andersen, P.H., 2014. An equine pain face.
Veterinary Anaesthesia and Analgesia 42, 103–114.
Graubner, C., Gerber, V., Doherr, M., Spadavecchia, C., 2011. Clinical application and
reliability of a post abdominal surgery pain assessment scale (PASPAS) in horses.
The Veterinary Journal 188, 178–183.
Grulke, S., Olle, E., Detilleux, J., Gangl, M., Caudron, I., Serteyn, D., 2001. Determination
of a gravity and shock score for prognosis in equine surgical colic. Journal of the
Veterinary Medical Association 48, 465–473.
Hansen, B.D., 2003. Assessment of pain in dogs: Veterinary clinical studies. ILAR
Journal 44, 197–205.
Hawker, G.A., Mian, S., Kendzerska, T., French, M., 2011. Measures of pain. Arthritis
Care & Research 63, S240–S252.
Hellyer, P.W., 2002. Treatment of pain in dogs and cats. Journal American Veterinary
Medical Association 221, 212–215.
Lindegaard, C., Thomsen, M.H., Larsen, S., Andersen, P.H., 2010. Analgesic efficacy
of intra-articular morphine in experimentally induced radiocarpal synovitis in
horses. Veterinary Anaesthesia and Analgesia 37, 171–185.
Malaty, H.M., Abudayyeh, S., O’Malley, K.J., Wilsey, M.J., Fraley, K., Gilger, M.A., Hollier,
D., Graham, D.Y., Rabeneck, L., 2005. Development of a multidimensional measure
for recurrent abdominal pain in children: Population-based studies in three
settings. Pediatrics 115, e209–e215.
Myles, P.S., 2007. Using the Bland-Altman method to measure agreement with
repeated measures. British Journal of Anaesthesia 99, 309–311.
Pritchett, L.C., Ulibarri, C., Roberts, M.C., Schneider, R.K., Sellon, D.C., 2003.
Identification of potential physiological and behavioural indicators of
postoperative pain in horses after exploratory celiotomy for colic. Applied Animal
Behaviour Science 80, 31–43.
Prunier, A., Mounier, L., Le Neidre, P., Leterrier, C., Mormède, P., Paulmier, V., Prunet,
P., Terlouw, C., Guatteo, R., 2013. Identifying and monitoring pain in farm animals:
A review. Animal: An International Journal of Animal Bioscience 7, 998–1010.
Robertson, S.A., 2002. What is pain? Journal American Veterinary Medical Association
221, 202–205.
Sotocinal, S.G., Sorge, R.E., Zaloum, A., Tuttle, A.H., Martin, L.J., Wieskopf, J.S.,
Mapplebeck, J.C., Wei, P., Zhan, S., Zhang, S., et al., 2011. The Rat Grimace Scale:
A partially automated method for quantifying pain in the laboratory rat via facial
expressions. Molecular Pain 7, 55.
Sutton, G.A., Dahan, R., Turner, D., Paltiel, O., 2013a. A behaviour-based pain scale
for horses with acute colic: Scale construction. The Veterinary Journal 196,
394–401.
Sutton, G.A., Paltiel, O., Soffer, M., Turner, D., 2013b. Validation of two behaviour-
based pain scales for horses with acute colic. The Veterinary Journal 197, 646–
650.
van Loon, J.P.A.M., Back, W., Hellebrekers, L.J., van Weeren, P.R., 2010. Application
of a composite pain scale to objectively monitor horses with somatic and visceral
pain under hospital conditions. Journal of Equine Veterinary Science 30, 641–
649.
van Loon, J.P.A.M., Jonckheer-Sheehy, V.S.M., Back, W., van Weeren, P.R., Hellebrekers,
L.J., 2014. Monitoring equine visceral pain with a composite pain scale score and
correlation with survival after emergency gastrointestinal surgery. The Veterinary
Journal 200, 109–115.
Valverde, A., Gunkel, C.I., 2005. Pain management in horses and farm animals. Journal
Veterinary Emergency and Critical Care 15, 295–307.
Wagner, A.E., 2010. Effects of stress on pain in horses and incorporating pain scales
for equine practice. Veterinary Clinics of North America. Equine Practice 26,
481–492.
White, N.A., Elward, A., Moga, K.S., Ward, D.L., Sampson, D.M., 2005. Use of web-based
data collection to evaluate analgesic administration and the decision for surgery
in horses with colic. Equine Veterinary Journal 37, 347–350.
364 J.P.A.M. van Loon, M.C. Van Dierendonck/The Veterinary Journal 206 (2015) 356–364