Advertisement

Investigating the effect of anxiety on pain scores in dogs

Published:November 09, 2021DOI:https://doi.org/10.1016/j.vaa.2021.07.005

      Abstract

      Objective

      To investigate the relationship between anxiety and pain scores using the Glasgow Composite Measure Pain Scale–Short Form (CMPS-SF) in dogs.

      Study

      Prospective observational study.

      Animals

      A group of 18 dogs undergoing surgical management of stifle disease.

      Methods

      Preoperatively dogs were scored using the CMPS-SF, the anxiety behaviour-based Reactivity Evaluation Form (REF), a Visual Analogue Scale (VAS) for anxiety and a sedation score. Assessments of pain, anxiety and sedation were repeated approximately 2–6 hours postoperatively. Dogs were divided into groups based on preoperative REF (‘Low REF’ and ‘High REF’), and VAS scores (‘Low VAS’ and ‘High VAS’). Scores (CMPS-SF, REF, VAS and sedation) were compared between groups using Mann–Whitney U tests. Preoperative and postoperative CMPS-SF, REF and VAS scores were compared using Wilcoxon signed-rank tests. Relationships between anxiety and CMPS-SF scores were assessed using a Spearman rank correlation coefficient. Scores are presented as median (range). A p value of < 0.05 was considered significant.

      Results

      When divided based on REF, CMPS-SF scores did not differ between groups preoperatively [Low REF: 2 (0–3), High REF: 2 (1–3); p = 0.509] or postoperatively [Low REF: 3 (2–5), High REF: 3 (2–5); p = 0.624]. When divided based on VAS, CMPS-SF scores did not differ between groups preoperatively [Low VAS: 2 (0–2), High VAS: 2 (1–3); p = 0.215] or postoperatively [Low VAS: 3 (2–5), High VAS: 3 (2–5); p = 1]. Postoperative REF [pre: 4.5 (2–8), post: 5 (4–10); p = 0.0105] and CMPS-SF scores [pre: 2 (0–3), post: 3 (2–5); p = 0.0318] increased significantly compared with preoperative scores.

      Conclusions and clinical relevance

      No apparent relationship exists between baseline anxiety levels and CMPS-SF scores. Understanding the influence of anxiety when using the CMPS-SF is important when assessing pain in dogs. Anxiety and pain may increase postoperatively in dogs undergoing orthopaedic surgery.

      Keywords

      Introduction

      The assessment of pain in veterinary patients remains a challenge. Given the subjective nature of the pain experience, and the inability of animals to report their pain, we rely on behavioural observations to guide pain management (
      • Mathews K.
      • Kronen P.W.
      • Lascelles D.
      • et al.
      Guidelines for recognition, assessment and treatment of pain.
      ). Despite the recent development of validated pain assessment tools in both dogs (
      • Reid J.
      • Nolan A.M.
      • Hughes J.M.L.
      • et al.
      Development of the short-form Glasgow Composite Measure Pain Scale (CMPS-SF) and derivation of an analgesic intervention score.
      ) and cats (
      • Brondani J.T.
      • Mama K.R.
      • Luna S.P.L.
      • et al.
      Validation of the English version of the UNESP-Botucatu multidimensional composite pain scale for assessing postoperative pain in cats.
      ;
      • Reid J.
      • Scott E.M.
      • Calvo G.
      • Nolan A.M.
      Definitive Glasgow acute pain scale for cats: validation and intervention level.
      ), several factors may confound these assessments. Scales are validated in a specific environment (
      • Buisman M.
      • Hasiuk M.M.
      • Gunn M.
      • Pang D.S.
      The influence of demeanour on scores from two validated feline pain assessment scales during the perioperative period.
      ), and their use may be influenced by age, breed and temperament (
      • Mathews K.
      • Kronen P.W.
      • Lascelles D.
      • et al.
      Guidelines for recognition, assessment and treatment of pain.
      ), as well as the effects of anaesthetic drugs (
      • Guillot M.
      • Rialland P.
      • Nadeau M.È.
      • et al.
      Pain induced by a minor medical procedure (Bone marrow aspiration) in Dogs: comparison of Pain scales in a pilot study.
      ;
      • Rialland P.
      • Authier S.
      • Guillot M.
      • et al.
      Validation of orthopedic postoperative pain assessment methods for dogs: a prospective, blinded, randomized, placebo-controlled study.
      ) and observer experience (
      • Barletta M.
      • Young C.N.
      • Quandt J.E.
      • Hofmeister E.H.
      Agreement between veterinary students and anesthesiologists regarding postoperative pain assessment in dogs.
      ). It seems obvious that any factor which may influence the behaviour of our patients has the potential to interfere with pain assessment.
      The Glasgow Composite Measure Pain Scale–Short Form (CMPS-SF) (
      • Reid J.
      • Nolan A.M.
      • Hughes J.M.L.
      • et al.
      Development of the short-form Glasgow Composite Measure Pain Scale (CMPS-SF) and derivation of an analgesic intervention score.
      ) is a behaviour-based composite pain scale, with a defined intervention level for provision of analgesia. It is valid in different environments (
      • Murrell J.C.
      • Psatha E.P.
      • Scott E.M.
      • et al.
      Application of a modified form of the Glasgow pain scale in a veterinary teaching centre in the Netherlands.
      ), and recently an Italian version has also been validated (
      • Della Rocca G.
      • Colpo R.
      • Reid J.
      • et al.
      Creation and validation of the Italian version of the Glasgow composite measure pain scale-short form (ICMPS-SF).
      ). It is widely used as the clinical standard for assessing acute pain in dogs (
      • Calvo G.
      • Holden E.
      • Reid J.
      • et al.
      Development of a behaviour-based measurement tool with defined intervention level for assessing acute pain in cats.
      ).
      Anxiety is an emotional response in anticipation of a stimulus or situation which an animal perceives as dangerous, leading to a somatic and behavioural stress response (
      • Palestrini C.
      • Minero M.
      • Cannas S.
      • et al.
      Efficacy of a diet containing caseinate hydrolysate on signs of stress in dogs.
      ). Many animals will show behavioural signs of fear or anxiety in a veterinary environment, the extent of which will depend on species, breed, previous experiences and individual variation (
      • Palestrini C.
      Situational sensitivities.
      ). Fear and anxiety can affect how pain is experienced by human patients (
      • Rhudy J.L.
      • Meagher M.W.
      Fear and anxiety: divergent effects on human pain thresholds.
      ), and it has been suggested that the use of behaviour-based pain scales in dogs can be influenced by the presence of anxiety (
      • Carsten R.E.
      • Hellyer P.W.
      • Bachand A.M.
      • LaRue S.M.
      Correlations between acute radiation scores and pain scores in canine radiation patients with cancer of the forelimb.
      ).
      • Buisman M.
      • Hasiuk M.M.
      • Gunn M.
      • Pang D.S.
      The influence of demeanour on scores from two validated feline pain assessment scales during the perioperative period.
      showed that demeanour in cats can affect the assessment of pain using certain validated pain scales. Certain personality traits such as extroversion and neuroticism may also affect the expression and experience of pain in humans (
      • Ramirez-Maestre C.
      • Martinez A.E.L.
      • Zarazaga R.E.
      Personality characteristics as differential variables of the pain experience.
      ), horses (
      • Ijichi C.
      • Collins L.M.
      • Elwood R.W.
      Pain expression is linked to personality in horses.
      ) and dogs (
      • Lush J.
      • Ijichi C.
      A preliminary investigation into personality and pain in dogs.
      ).
      The effect of anxiety is often accepted to limit the usefulness of physiological measures of pain (
      • Mathews K.
      • Kronen P.W.
      • Lascelles D.
      • et al.
      Guidelines for recognition, assessment and treatment of pain.
      ), but its impact on the use of behaviour-based pain assessment tools is still unclear. The aim of this study was to investigate the effect of anxiety on the use of CMPS-SF for assessment of pain in hospitalized canine patients undergoing surgery for management of stifle disease. We hypothesized that anxious dogs would be assigned higher pain scores than non-anxious dogs undergoing similar surgical procedures.

      Materials and methods

      This study was performed at the Small Animal Hospital, University of Glasgow, following ethical approval by the School of Veterinary Medicine Ethics and Welfare Committee (Reference EA07/19). Dogs undergoing general anaesthesia for surgical management of stifle disease were enrolled, with informed owner consent obtained for inclusion in the study. Dogs showing aggressive behaviour that would prevent safe assessments of pain and anxiety, dogs undergoing bilateral surgery, or dogs undergoing any other procedures under the same anaesthetic (excluding preoperative and/or postoperative radiography) were excluded.
      A sample size estimation was performed using a Bland–Altman nomogram as described in
      • Petrie A.
      • Watson P.
      Statistics for Veterinary and Animal Science.
      for comparison of two independent groups. Based on the findings by
      • Murrell J.C.
      • Psatha E.P.
      • Scott E.M.
      • et al.
      Application of a modified form of the Glasgow pain scale in a veterinary teaching centre in the Netherlands.
      that the difference in mean CMPS-SF pain scores of dogs experiencing mild or moderate pain 6 hours postoperatively was 1.937, a difference in pain scores of 2 between groups was considered clinically relevant. A sample size of 18 dogs (nine in each group) was estimated to have 80% power at a significance level of 0.05, assuming a standard deviation in pain scores of 1.5 (reported by
      • Drygras K.A.
      • McClure S.R.
      • Goring R.L.
      • et al.
      Effect of cold compression therapy on postoperative pain, swelling, range of motion, and lameness after tibial plateau leveling osteotomy in dogs.
      ).
      After admission to the hospital, a preoperative baseline assessment was performed on each dog, following a period of at least 30 minutes to acclimatize to the hospital environment, and prior to the onset of action of any premedication agents. All assessments were performed by the same observer in real time. Pain assessment was performed using the CMPS-SF (
      • Reid J.
      • Nolan A.M.
      • Hughes J.M.L.
      • et al.
      Development of the short-form Glasgow Composite Measure Pain Scale (CMPS-SF) and derivation of an analgesic intervention score.
      ). An assessment of anxiety was performed using a behaviour-based scoring system, the Reactivity Evaluation Form (REF) (
      • Palestrini C.
      • Minero M.
      • Cannas S.
      • et al.
      Efficacy of a diet containing caseinate hydrolysate on signs of stress in dogs.
      ). This scoring system involves first assessing a dog outside the kennel and assigning a score based on its behaviour towards the investigator, which is then repeated after entering the kennel without interacting with the dog. Dogs are also assigned scores based on the presence or absence of certain behaviours, and the final REF score is the sum of these scores, with a higher score representing a greater level of anxiety (Appendix A). A subjective anxiety assessment was also performed using a Visual Analogue Scale (VAS), where 0 mm represented no anxiety and 100 mm represented the highest possible level of anxiety. Assessments of pain and anxiety were performed at the same time, with the dog first observed from outside the kennel. The observer then entered the kennel and interacted with the dog as required by the scoring systems to complete the assessment. A simple descriptive scale was used to assess sedation, as previously described by
      • Murrell J.C.
      • Psatha E.P.
      • Scott E.M.
      • et al.
      Application of a modified form of the Glasgow pain scale in a veterinary teaching centre in the Netherlands.
      , with possible scores of 0 (‘fully alert, able to stand and walk’), 1 (‘alert, able to maintain sternal recumbency and walk but may be ataxic’), 2 (‘drowsy, able to maintain sternal recumbency but unable to stand’) or 3 (‘fast asleep, unable to raise head’).
      If a dog was assigned a pain score ≥ 6/24, the validated intervention point for the CMPS-SF (
      • Reid J.
      • Nolan A.M.
      • Hughes J.M.L.
      • et al.
      Development of the short-form Glasgow Composite Measure Pain Scale (CMPS-SF) and derivation of an analgesic intervention score.
      ), the veterinary surgeon in charge of the case was informed to allow additional analgesia to be prescribed. If the assessment of pain or demeanour was thought to be causing distress to the animal, the procedure was stopped immediately.
      Each dog then underwent the planned surgical procedure. The anaesthetic protocol and surgical procedure were decided by the responsible veterinary surgeon in each case. Postoperative analgesia was decided as appropriate for the case (e.g., methadone and non-steroidal anti-inflammatory drugs). A postoperative assessment of pain, anxiety and sedation was performed once the animal had recovered sufficiently from general anaesthesia (approximately 2–6 hours postoperatively). Level of recovery from anaesthesia was considered appropriate if the animal was responsive, mobile and able to interact with the observer to allow assessments of pain and anxiety to be performed. The age, breed, sex and procedure undertaken were also recorded for each dog. Premedication, induction, maintenance anaesthetic agents and duration of anaesthesia were identified from the anaesthetic record for each case, together with intraoperative analgesia and any additional sedation.
      When data collection was completed, dogs were separated into groups based on their preoperative assessments of anxiety. As neither measure used has a validated ‘cut-off’ point, the 50% of dogs scoring highest REF scores were designated ‘High REF’ and the lowest scoring 50% of dogs designated ‘Low REF’. A similar division was made based on VAS scores: ‘Low VAS’ and ‘High VAS’.

      Statistical analysis

      Anaesthetic time was tested for normality using the chi-square goodness of fit test and is presented as mean (± standard deviation). The Student t test was used to compare anaesthetic times between groups. Ordinal data (VAS, REF and CMPS-SF) were assumed not to be normally distributed and are presented as median (range). The Wilcoxon signed-rank test was used to compare postoperative VAS, REF and CMPS-SF scores with preoperative scores. The Mann–Whitney U test was used to compare CMPS-SF, REF, VAS and sedation scores between groups. The type of premedication used was compared between groups using the chi-square test. The incidence of sedative administration before or during anaesthetic recovery was compared between groups using Fisher’s exact test. The relationships between preoperative anxiety scores and CMPS-SF scores were assessed using Spearman’s rank correlation coefficient. The change in CMPS-SF scores was calculated as the preoperative score subtracted from the postoperative score, and the relationship between preoperative anxiety scores and change in CMPS-SF scores was assessed using Spearman’s rank correlation coefficient. A p value of < 0.05 was considered significant.
      Descriptive statistical analysis, chi-square tests, Fisher’s exact test and Student t test were performed using Microsoft Excel (Excel Version 2002; Microsoft Corporation, WA, USA). Online calculators were used to calculate Mann–Whitney U test (
      • Social Science Statistics
      Mann-Whitney U Test Calculator.
      ), Wilcoxon signed-rank test (
      • Social Science Statistics
      Wilcoxon Signed-Ranks Test Calculator.
      ) and Spearman’s rank correlation coefficients (
      • Social Science Statistics
      Spearman’s Rho Calculator.
      ).

      Results

      A total of 18 dogs were included, with a median age of 5.72 years (0.57–9.86). All dogs underwent surgical procedures for management of stifle disease. Demographic data are summarized in Table 1. Consent was obtained from owners of 22 dogs, with four dogs not proceeding to inclusion in the study owing to researcher unavailability. A dog was initially excluded after baseline assessment owing to cancellation of the planned surgery. This dog was included in the study at a later date, with a repeat baseline assessment performed when re-admitted to hospital for the rescheduled procedure.
      Table 1The total number of dogs included in the study of each breed, sex and procedure. Numbers are also shown for each group when divided based on preoperative Reactivity Evaluation Form (REF) and Visual Analogue Scale (VAS) scores
      BreedTotal number of dogsNumber of dogs in each group
      Low REFHigh REFLow VASHigh VAS
      Bichon Frise10101
      Chihuahua10101
      Cocker Spaniel11010
      Crossbreed11010
      Doberman Pinscher10101
      English Bulldog10101
      French Bulldog21120
      Labrador74343
      Springer Spaniel10101
      West Highland White Terrier22011
      Sex
      Female entire21120
      Female ovariohysterectomized93636
      Male entire21111
      Male castrated54132
      Procedure
      Tibial plateau levelling osteotomy (TPLO)127566
      Medial patellar luxation repair (MPL)20211
      Extracapsular suture10101
      Tibial tuberosity fracture repair11010
      Femoral osteotomy11010
      TPLO and MPL10101
      Anaesthetic records were available for 17 out of 18 dogs. As premedication, nine dogs were administered acepromazine (5.2–30 μg kg–1) and methadone (0.19–0.31 mg kg–1), five dogs were administered medetomidine (3.0–8.1 μg kg–1) and methadone (0.19–0.3 mg kg–1), one dog was administered acepromazine (19 μg kg–1), medetomidine (3.8 μg kg–1) and methadone (0.27 mg kg–1), and one dog was administered medetomidine (7.5 μg kg–1) and morphine (0.28 mg kg–1). Anaesthesia was induced with propofol (n = 14) or alfaxalone (n = 2), and in one case the induction agent used was not recorded. Anaesthesia was maintained with isoflurane (n = 17), although in two cases, the inhalant anaesthetic was changed from isoflurane to sevoflurane (or vice versa) during anaesthesia. Nitrous oxide was also used in six dogs. Local anaesthetic blocks were performed in 16 cases (femoral and sciatic blocks: n = 15; sciatic block only: n = 1). Intraoperative analgesia used included non-steroidal anti-inflammatory drugs (n = 8), paracetamol (n = 4), methadone (n = 9), morphine (n = 1), fentanyl (n = 7) and ketamine (n =1). Effectiveness of the local anaesthetic block was not directly assessed. Total anaesthetic time was 209 minutes (±55) and was normally distributed (p = 0.57). Anaesthetic time was not recorded in one case.
      Acepromazine was administered to one dog during anaesthesia (5.4 μg kg–1) on two occasions, approximately 165 and 110 minutes prior to the end of anaesthesia. In the immediate postanaesthetic period, one dog was sedated with 1.8 μg kg–1 medetomidine intravenously.
      All dogs tolerated the assessments of pain, anxiety and sedation well. All dogs were able to walk at the time of postoperative assessment; therefore all pain scores were out of 24. No dogs scored a pain score of ≥ 6/24, and no rescue analgesia was required.
      Based on preoperative REF scores, dogs were divided into two groups, ‘Low REF’ and ‘High REF’. Preoperative REF scores were significantly different between groups (Fig. 1). They were also divided into two groups based on preoperative VAS scores, ‘Low VAS’ and ‘High VAS’. Preoperative VAS scores were also significantly different between groups (Fig. 2). Anaesthetic time was not significantly different between ‘Low REF’ [214 minutes (± 57)] and ‘High REF’ [206 minutes (± 55)] (p = 0.729), nor between ‘Low VAS’ [203 minutes (± 49)] and ‘High VAS’ [209 minutes (± 54)] (p = 0.834). The type of premedication administered (acepromazine and opioid, α2-adrenoceptor agonist and opioid, or acepromazine, α2-adrenoceptor agonist and opioid) was independent of both REF (p = 0.549) and VAS (p = 0.549) groups. Both dogs which were administered additional sedation before or during recovery were in ‘Low REF’ and ‘Low VAS’ groups; however, the difference in proportions was not statistically significant between ‘Low REF’ and ‘High REF’ groups (p = 0.206), nor between ‘Low VAS’ and ‘High VAS’ groups (p = 0.206).
      Figure 1
      Figure 1A box and whisker plot showing preoperative and postoperative Reactivity Evaluation Form (REF) scores in 18 dogs undergoing surgery for management of stifle disease, divided into groups based on preoperative REF scores. Dark grey represents ‘Low REF’ (n = 9), light grey represents ‘High REF’ (n = 9). The box represents the interquartile range, with the central line representing the median value. Whiskers represent minimum and maximum values. Dots represent outliers (>1.5 × interquartile range above or below the first or third quartile, respectively). Preoperative scores were significantly different between ‘Low REF’ and ‘High REF’ groups (p = 0.0004). Overall postoperative REF scores were significantly increased compared with preoperative scores (p = 0.0105). When divided into ‘Low REF’ and ‘High REF’ groups, there was a significant increase in REF scores postoperatively in the ‘Low REF’ group (p < 0.05), but not in the ‘High REF’ group (p > 0.05). In the ‘Low REF’ group, the preoperative median and upper quartile are the same value, as are the postoperative median and lower quartile.
      Figure 2
      Figure 2Preoperative and postoperative Visual Analogue Scale (VAS) scores in 18 dogs undergoing surgery for management of stifle disease, divided into groups based on preoperative VAS scores. Dark grey represents ‘Low VAS’ (n = 9), light grey represents ‘High VAS’ (n = 9). The box represents the interquartile range, with the central line representing the median value. Whiskers represent minimum and maximum values. Preoperative scores were significantly different between ‘Low VAS’ and ‘High VAS’ groups (p = 0.0004). Overall, there was no significant difference between preoperative and postoperative VAS scores (p = 0.603). Postoperative VAS scores were increased significantly compared with preoperative scores in the ‘Low VAS’ group (p < 0.05), but not in the ‘High VAS’ group (p > 0.05).
      All preoperative sedation scores were 0. At the time of postoperative assessment, 10/18 dogs were assigned a sedation score of 0, with the remainder (8/18) assigned a sedation score of 1. When postoperative sedation scores were compared between groups, there was no statistically significant difference between groups when divided based on REF [Low REF: 0 (0–1), High REF: 0 (0–1); p = 0.968], or VAS [Low VAS: 0 (0–1), High VAS: 1 (0–1); p = 0.453].
      There was no statistically significant difference in VAS scores between preoperative and postoperative assessments overall. However, when divided into ‘Low VAS’ and ‘High VAS’ groups based on preoperative scores, postoperative VAS scores increased significantly in the ‘Low VAS’ groups compared with preoperative scores. In the ‘High VAS’ groups, there was no significant difference between preoperative and postoperative scores (Fig. 2).
      Overall, postoperative REF scores were significantly higher than preoperative scores. When divided into ‘Low REF’ and ‘High REF’ groups based on preoperative scores, postoperative REF scores were significantly higher than preoperative scores in the ‘Low REF group’. This difference was not seen in the ‘High REF’ group, with no significant difference between preoperative and postoperative scores (Fig. 1).
      Postoperatively, CMPS-SF scores [3 (2–5)] were significantly higher than preoperative scores [2 (0–3)] (p = 0.0032). When divided into groups based on REF, CMPS-SF scores did not differ between groups preoperatively or postoperatively (Fig. 3). There was no significant correlation between preoperative REF and CMPS-SF scores (rs = 0.0551, p = 0.828), nor between preoperative REF scores and postoperative CMPS-SF scores (rs = 0.142, p = 0.574). The relationship between preoperative REF scores and the change in CMPS-SF scores was also not significant (rs = 0.0401, p = 0.874).
      Figure 3
      Figure 3Preoperative and postoperative Glasgow Composite Measure Pain Scale–Short Form (CMPS-SF) scores in 18 dogs undergoing surgery for management of stifle disease, divided into groups based on preoperative Reactivity Evaluation Form scores (REF). Dark grey represents ‘Low REF’ (n = 9), light grey represents ‘High REF’ (n = 9). The box represents the interquartile range, with the central line representing the median value. Whiskers represent minimum and maximum values. There was no significant difference in CMPS-SF scores between groups neither preoperatively (p= 0.509) nor postoperatively (p = 0.642). In the ‘Low REF’ group, the preoperative median and upper quartile are the same value.
      Similarly, when divided into groups based on VAS, there was no significant difference in CMPS-SF scores between ‘Low VAS’ and ‘High VAS’ groups preoperatively or postoperatively (Fig. 4). The relationship between preoperative VAS and CMPS-SF scores was not significant (rs = 0.341, p = 0.167), and there was no significant correlation between preoperative VAS scores and postoperative CMPS-SF scores (rs = 0.116, p = 0.646). The relationship between preoperative VAS scores and the change in CMPS-SF scores was also not significant (rs = –0.117, p = 0.644).
      Figure 4
      Figure 4Preoperative and postoperative Glasgow Composite Measure Pain Scale–Short Form (CMPS-SF) scores in 18 dogs undergoing surgery for management of stifle disease, divided into groups based on preoperative Visual Analogue Scale (VAS) scores. Dark grey represents ‘Low VAS’ (n = 9), light grey represents ‘High VAS’ (n = 9). The box represents the interquartile range, with the central line representing the median value. Whiskers represent minimum and maximum values. There was no significant difference in CMPS-SF scores between groups either preoperatively (p = 0.215) or postoperatively (p = 1). In the ‘Low VAS’ group, the preoperative median and upper quartile are the same value.

      Discussion

      Our results show no relationship between preoperative anxiety levels and preoperative or postoperative CMPS-SF scores. The higher REF scores seen postoperatively suggest that anxiety levels increase following surgery.
      Complicated interactions exist between stress-related behaviour and pain. Given the unpleasant sensory and emotional experience associated with pain (
      International Association for the Study of Pain (IASP)
      IASP Terminology.
      ), pain will likely contribute to the stress an animal experiences in a veterinary environment (
      • Lind A.K.
      • Hydbring-Sandberg E.
      • Forkamn B.
      • Keeling L.J.
      Assessing stress in dogs during a visit to the veterinary clinic: Correlations between dog behaviour in standardized tests and assessments by veterinary staff and owners.
      ). Indeed, section D of the CMPS-SF requires an assessment of demeanour, with one option being ‘nervous, anxious or fearful’ (
      • Reid J.
      • Nolan A.M.
      • Hughes J.M.L.
      • et al.
      Development of the short-form Glasgow Composite Measure Pain Scale (CMPS-SF) and derivation of an analgesic intervention score.
      ). However, pain is not the only factor contributing to anxiety during hospitalization and a novel environment, previous negative experiences, or restraint may all contribute (
      • Palestrini C.
      Situational sensitivities.
      ).
      Given the requirement to assess demeanour when using the CMPS-SF, it would be easy to conclude that dogs that are already anxious in a veterinary environment will automatically score highly on this scale.
      • Carsten R.E.
      • Hellyer P.W.
      • Bachand A.M.
      • LaRue S.M.
      Correlations between acute radiation scores and pain scores in canine radiation patients with cancer of the forelimb.
      observed that CMPS-SF scores in dogs undergoing radiotherapy decreased over time and suggested that as animals became acclimatized to the clinic environment, the likelihood of assessing them as ‘nervous, anxious or fearful’ diminished.
      However, if the behaviours assessed by the CMPS-SF are specific to pain, the presence of anxiety attributed to other factors may not matter, as other pain associated behaviours must be present for a dog to reach the validated intervention level. A study investigating various behavioural assessments of stress in a veterinary clinic did not find any correlation between pain assessments (using a subjective Likert Scale) and the behaviour tests carried out (
      • Lind A.K.
      • Hydbring-Sandberg E.
      • Forkamn B.
      • Keeling L.J.
      Assessing stress in dogs during a visit to the veterinary clinic: Correlations between dog behaviour in standardized tests and assessments by veterinary staff and owners.
      ), which could suggest that pain behaviours and behaviours associated with stress can be distinguished. Given the lack of correlations between assessments of anxiety and CMPS-SF scores with no significant difference in pain scores between high anxiety and low anxiety groups, our results suggest no relationship between baseline anxiety scores and CMPS-SF.
      Our findings contrast with those of previous studies, which suggest that demeanour and personality may have an impact on pain assessment and resultant scores in animals.
      • Buisman M.
      • Hasiuk M.M.
      • Gunn M.
      • Pang D.S.
      The influence of demeanour on scores from two validated feline pain assessment scales during the perioperative period.
      found that aggressive and shy cats were assigned higher scores on certain validated pain scales.
      • Lush J.
      • Ijichi C.
      A preliminary investigation into personality and pain in dogs.
      suggested that dogs with an extrovert personality express pain more clearly. In horses, neurotic and extrovert personality traits may influence the behavioural expression of pain (
      • Ijichi C.
      • Collins L.M.
      • Elwood R.W.
      Pain expression is linked to personality in horses.
      ). In humans, we know that the presence of fear or anxiety will affect how an individual experiences pain. The presence of anxiety decreases withdrawal thresholds to radiant heat, whereas the presence of fear increases pain thresholds (
      • Rhudy J.L.
      • Meagher M.W.
      Fear and anxiety: divergent effects on human pain thresholds.
      ).
      There are numerous reasons why our findings differ from those of previous studies. Given the need to assess both spontaneous and interactive behaviours to accurately assess pain (
      • Reid J.
      • Nolan A.M.
      • Hughes J.M.L.
      • et al.
      Development of the short-form Glasgow Composite Measure Pain Scale (CMPS-SF) and derivation of an analgesic intervention score.
      ), patient interaction may be important when distinguishing pain and personality. For example,
      • Lush J.
      • Ijichi C.
      A preliminary investigation into personality and pain in dogs.
      found that dogs with extrovert personalities were assigned higher pain scores using CMPS-SF. However, in this study, analysis of mobility and palpation of wounds (sections B and C of the scale) were omitted, as scoring was performed retrospectively using videos. It is possible that the effect of personality and pain assessment may be more pronounced when animal interactions are excluded.
      In addition, the behavioural expression of pain is species specific (
      • Mathews K.
      • Kronen P.W.
      • Lascelles D.
      • et al.
      Guidelines for recognition, assessment and treatment of pain.
      ) and we cannot assume that factors affecting the assessment of pain in one species will automatically be applicable to another.
      There is also a difference between personality and anxiety. Personality is defined as ‘behavioural consistency across times and contexts’ (
      • Miklósi Á.
      • Turcsán B.
      • Kubinyi E.
      The personality of dogs.
      ), whereas anxious behaviour is a response in anticipation of a specific stimulus or situation (
      • Palestrini C.
      Situational sensitivities.
      ). Certain personality types may be associated with anxious behaviour. For example, neurotic personality traits include a predisposition to anxiety, increased sensitivity to stress and lack of emotional stability (
      • Ijichi C.
      • Collins L.M.
      • Elwood R.W.
      Pain expression is linked to personality in horses.
      ). However, each individual’s anxiety probably varies depending on context.
      Our study also found that postoperative REF scores were significantly higher than preoperative scores, indicating that anxiety may increase postoperatively. This overall increase appeared to be primarily the result of an increase in scores in the ‘Low REF’ group, which was not seen in the ‘High REF’ group. Similarly, a statistically significant increase in postoperative VAS scores was seen in the ‘Low VAS’ group, but not in the ‘High VAS’ group. This suggests that the postoperative increase in anxiety is greater in those with a lower baseline anxiety level. There is evidence to suggest that perioperative stress occurs in veterinary patients:
      • Väisänen M.A.
      • Valros A.E.
      • Hakoaja E.
      • et al.
      Pre-operative stress in dogs – a preliminary investigation of behaviour and heart rate variability in healthy hospitalised dogs.
      found that 80% of dogs hospitalized for ovariohysterectomy showed panting and displacement behaviours preoperatively. Following ovariohysterectomy, plasma cortisol levels increase from baseline, with behavioural changes and changes in other haematological and biochemical markers, indicating a physiological stress response to factors such as surgery, pain, dysphoria, handling and hospitalization (
      • Siracusa C.
      • Manteca X.
      • Cerón J.
      • et al.
      Perioperative stress response in dogs undergoing elective surgery: variations in behavioural, neuroendocrine, immune and acute phase responses.
      ). The behavioural changes seen in our study indicate that dogs may well experience an emotional response to this perioperative stress and highlights the importance of minimizing stress and anxiety in our patients in the perioperative period. This change in anxiety may be easier to detect in dogs with lower baseline anxiety levels. Anxious dogs may also experience greater anxiety in the postoperative period, but specific changes in behaviour may be less obvious.
      Possible limitations of this study need to be considered. The same observer assessed pain, anxiety and sedation, and was not blinded to the surgical procedure undertaken and medication administered. This may have subconsciously influenced the scores assigned to each animal. However, the use of a single observer assessing dogs in real time eliminates interobserver variability. The sample size investigated was small; however, the changes in REF and VAS scores showed statistical significance, suggesting adequate sample size.
      Animals in this study were assigned low pain scores, with none exceeding the validated intervention level. Reassuringly, this shows that the analgesia used was effective in controlling postoperative pain but may have reduced the range of pain behaviours observed (
      • Murrell J.C.
      • Psatha E.P.
      • Scott E.M.
      • et al.
      Application of a modified form of the Glasgow pain scale in a veterinary teaching centre in the Netherlands.
      ). However, the inclusion of a negative control group where analgesia was withheld would raise ethical issues.
      The REF scoring system used to assess anxiety is not validated. A subjective VAS scale was also used to assess anxiety. Neither scale has a validated ‘cut-off’ point for determining whether a dog is anxious or not; therefore, our group division was arbitrarily made with the 50% of animals with the lowest scores in the ‘low’ group and the 50% with the highest scores into the ‘high’ group. Validating behavioural assessment tools is challenging owing to a lack of a gold standard and the requirement to interpret behavioural signs of an animal’s experience. However, it is possible for subjective observer ratings to be valid and reliable when assessing animal behaviour (
      • Meagher R.K.
      Observer ratings: validity and value as a tool for animal welfare research.
      ).
      Anaesthetic and surgical procedures were not standardized between dogs. Anaesthetic drugs may affect patient behaviour, and there is evidence that the presence of sedation may interfere with the use of CMPS-SF (
      • Guillot M.
      • Rialland P.
      • Nadeau M.È.
      • et al.
      Pain induced by a minor medical procedure (Bone marrow aspiration) in Dogs: comparison of Pain scales in a pilot study.
      ). Postoperatively some sedation was apparent, although no dogs were assigned a score greater than 1 on the sedation scale. Sedation scores, premedication and use of sedation before and during anaesthetic recovery were not significantly different between groups, so probably did not affect our results. The surgical procedures were not completely standardized; however, all were pelvic limb procedures. As the aim was to establish changes in pain score rather than looking at the pain/analgesia per se, different procedures were included. It was felt that all dogs would undergo a similar degree of tissue trauma and pain, with the population we studied representative of a typical caseload in orthopaedic referral practice.

      Conclusions

      Even when using validated pain scales, factors which may interfere with their use should be considered. However, our findings do not suggest a relationship between baseline anxiety and pain scores assigned using the CMPS-SF. It should not be automatically assumed that a pain score is artificially elevated simply because an animal is anxious. A number of factors, including pain, may contribute to the anxiety experienced by hospitalized veterinary patients, and anxiety may increase in our patients in the postoperative period.

      Acknowledgements

      The authors would like to thank the surgery department at the Small Animal Hospital, University of Glasgow, for their assistance in recruiting dogs for this study.

      Authors’ contributions

      BE: study design, data collection, statistical analysis, preparation of manuscript. PJM: study design, preparation of manuscript.

      Conflict of interest statement

      The authors declare no conflict of interest.

      Appendix A.

      Tabled 1
      Score
      Phase 1

      Investigator is standing outside the box
      Dog is seeking for contact, stands permanently against the fence1
      Dog is successively standing against the fence and slightly retreating2
      Dog is permanently retreating but presents some attempts in seeking contacts3
      Dog is prostrated in the back of the box, standing as far as possible from the fence, no attempts in seeking contacts4
      Phase 2

      Investigator comes inside the box and stays static without interacting with the dog
      Dog is seeking for physical contact with the investigator1
      Dog ends up by having a physical contact with the investigator after a few attempts2
      Dog attempts to approach the investigator but does not dare to have a physical contact3
      Dog is prostrated in the back of the box, standing as far as possible from the fence, no attempts in seeking contacts4
      Items
      FreezingYes = 1No = 0
      TremblingYes = 1No = 0
      YawningYes = 1No = 0
      Lip lickingYes = 1No = 0
      CrouchedYes = 1No = 0
      Tail between legsYes = 1No = 0
      Ears downYes = 1No = 0
      Defecation/urinationYes = 1No = 0
      Reprinted from Palestrini C, Minero M, Cannas S et al. (2010) Efficacy of a diet containing caseinate hydrolysate on signs of stress in dogs. J Vet Behav 5,309-317. © Copyright 2010 with permission from Elsevier.

      References

        • Barletta M.
        • Young C.N.
        • Quandt J.E.
        • Hofmeister E.H.
        Agreement between veterinary students and anesthesiologists regarding postoperative pain assessment in dogs.
        Vet Anaesth Anal. 2016; 43: 91-98
        • Brondani J.T.
        • Mama K.R.
        • Luna S.P.L.
        • et al.
        Validation of the English version of the UNESP-Botucatu multidimensional composite pain scale for assessing postoperative pain in cats.
        BMC Vet Res. 2013; 9: 143
        • Buisman M.
        • Hasiuk M.M.
        • Gunn M.
        • Pang D.S.
        The influence of demeanour on scores from two validated feline pain assessment scales during the perioperative period.
        Vet Anaest Anal. 2017; 44: 646-655
        • Calvo G.
        • Holden E.
        • Reid J.
        • et al.
        Development of a behaviour-based measurement tool with defined intervention level for assessing acute pain in cats.
        J Small Anim Pract. 2014; 55 (633–629)
        • Carsten R.E.
        • Hellyer P.W.
        • Bachand A.M.
        • LaRue S.M.
        Correlations between acute radiation scores and pain scores in canine radiation patients with cancer of the forelimb.
        Vet Anaes Analg. 2008; 35: 355-362
        • Della Rocca G.
        • Colpo R.
        • Reid J.
        • et al.
        Creation and validation of the Italian version of the Glasgow composite measure pain scale-short form (ICMPS-SF).
        Vet Ital. 2018; 54: 251-260
        • Drygras K.A.
        • McClure S.R.
        • Goring R.L.
        • et al.
        Effect of cold compression therapy on postoperative pain, swelling, range of motion, and lameness after tibial plateau leveling osteotomy in dogs.
        J Am Vet Med Assoc. 2011; 238: 1284-1291
        • Guillot M.
        • Rialland P.
        • Nadeau M.È.
        • et al.
        Pain induced by a minor medical procedure (Bone marrow aspiration) in Dogs: comparison of Pain scales in a pilot study.
        J Vet Intern Med. 2011; 25: 1050-1056
        • Ijichi C.
        • Collins L.M.
        • Elwood R.W.
        Pain expression is linked to personality in horses.
        Appl Anim Behav Sci. 2014; 152: 38-43
        • International Association for the Study of Pain (IASP)
        IASP Terminology.
        2017
        • Lind A.K.
        • Hydbring-Sandberg E.
        • Forkamn B.
        • Keeling L.J.
        Assessing stress in dogs during a visit to the veterinary clinic: Correlations between dog behaviour in standardized tests and assessments by veterinary staff and owners.
        J Vet Behav. 2017; 17: 24-31
        • Lush J.
        • Ijichi C.
        A preliminary investigation into personality and pain in dogs.
        J Vet Behav. 2018; 24: 62-68
        • Mathews K.
        • Kronen P.W.
        • Lascelles D.
        • et al.
        Guidelines for recognition, assessment and treatment of pain.
        J Small Anim Pract. 2014; 55: E10-E68
        • Meagher R.K.
        Observer ratings: validity and value as a tool for animal welfare research.
        Appl Anim Behav Sci. 2009; 119: 1-14
        • Miklósi Á.
        • Turcsán B.
        • Kubinyi E.
        The personality of dogs.
        in: Kaminski J. Marshall-Pescini S. The Social Dog: Behaviour and Cognition. Academic Press, San Diego, USA2014: 191-222
        • Murrell J.C.
        • Psatha E.P.
        • Scott E.M.
        • et al.
        Application of a modified form of the Glasgow pain scale in a veterinary teaching centre in the Netherlands.
        Vet Rec. 2008; 162: 403-408
        • Palestrini C.
        Situational sensitivities.
        in: Horwitz D.F. Mills D.S. BSAVA Manual of Canine and Feline Behavioural Medicine. 2nd edn. BSAVA, UK2009: 169-181
        • Palestrini C.
        • Minero M.
        • Cannas S.
        • et al.
        Efficacy of a diet containing caseinate hydrolysate on signs of stress in dogs.
        J Vet Behav. 2010; 5: 309-317
        • Petrie A.
        • Watson P.
        Statistics for Veterinary and Animal Science.
        3rd edn. John Wiley & Sons, UK2013: 184-187
        • Ramirez-Maestre C.
        • Martinez A.E.L.
        • Zarazaga R.E.
        Personality characteristics as differential variables of the pain experience.
        J Behav Med. 2004; 27: 147-165
        • Reid J.
        • Nolan A.M.
        • Hughes J.M.L.
        • et al.
        Development of the short-form Glasgow Composite Measure Pain Scale (CMPS-SF) and derivation of an analgesic intervention score.
        Anim Welf. 2007; 16: 97-104
        • Reid J.
        • Scott E.M.
        • Calvo G.
        • Nolan A.M.
        Definitive Glasgow acute pain scale for cats: validation and intervention level.
        Vet Rec. 2017; 180: 449
        • Rhudy J.L.
        • Meagher M.W.
        Fear and anxiety: divergent effects on human pain thresholds.
        Pain. 2000; 84: 65-75
        • Rialland P.
        • Authier S.
        • Guillot M.
        • et al.
        Validation of orthopedic postoperative pain assessment methods for dogs: a prospective, blinded, randomized, placebo-controlled study.
        PLoS One. 2012; 7: E49480
        • Siracusa C.
        • Manteca X.
        • Cerón J.
        • et al.
        Perioperative stress response in dogs undergoing elective surgery: variations in behavioural, neuroendocrine, immune and acute phase responses.
        Anim Welf. 2008; 17: 259-273
        • Social Science Statistics
        Mann-Whitney U Test Calculator.
        2021
        • Social Science Statistics
        Wilcoxon Signed-Ranks Test Calculator.
        2021
        • Social Science Statistics
        Spearman’s Rho Calculator.
        2021
        • Väisänen M.A.
        • Valros A.E.
        • Hakoaja E.
        • et al.
        Pre-operative stress in dogs – a preliminary investigation of behaviour and heart rate variability in healthy hospitalised dogs.
        Vet Anaesth Analg. 2005; 32: 158-167