Effects of feline personality on pain assessment with two validated behaviour-based scales

Abstract

Objectives

The aim of this study was to determine whether feline personality would affect the scores obtained with the Glasgow Feline Composite Measure Pain Scale (CMPS–Feline) and the Feline Grimace Scale (FGS), as well as the mechanical thresholds (MTs), both pre- and postoperatively.

Methods

Cats scheduled for elective neutering were included in the study. After at least 12 h of acclimatisation, the cats’ personalities were classified, based on the Feline-ality Assessment, as either ‘independent’ or ‘social’, always by the same investigator. Then, a second independent investigator assigned FGS and CMPS–Feline scores and measured the MTs with a mechanical algometer before (baseline) and after elective surgical neutering. Either a t-test or Mann–Whitney statistics, depending on data distribution, was used to compare continuous variables between independent and social cats.

Results

A total of 29 cats undergoing either orchiectomy or ovariectomy were included in the study. Of them, 14 were classified as independent and 15 as social. Preoperatively, the independent cats had higher CMPS–Feline (2.8 ± 1.6; P = 0.014) and FGS (0, range 2–1; P = 0.025) scores than the social cats (1.3 ± 1.2 and 0, range 0–0, respectively). Postoperatively, CMPS–Feline (P = 0.011) and FGS (P = 0.009) scores were higher in the independent cats (3, range 3–5 and 2, range 1–3, respectively) than in the social cats (2, range 1–3 and 1, range 0–1, respectively). MTs did not differ between personality groups either preoperatively or postoperatively. Within personality groups, there were no differences between pre- and postoperative study variables.

Conclusions and relevance

Feline personality affects the reliability of both the CMPS–Feline and FGS. This should be considered when these scales are used to assess perioperative pain in clinical cats and to titrate analgesic drugs to effect.

Plain language summary

Effects of feline personality on pain assessment with two validated behaviour-based scales

The objective of this study was to determine whether feline personality, classified as either ‘independent’ or ‘social’, would affect the scores obtained with two validated pain scales (FGS and CMPS–Feline) in cats undergoing surgical neutering. A total of 29 cats were included in the study. Their personalities were assessed with Feline-ality assessment, a validated tool, after at least 12 h of acclimatisation in the hospital, always by the same investigator. Then, a second independent investigator assigned FGS and CMPS before (baseline) and after surgery. Either a t-test or Mann–Whitney statistics, depending on data distribution, was used to compare continuous variables between independent and social cats. Of 29 cats, 14 were classified as ‘independent’ and 15 as ‘social’. Before surgery, the independent cats had higher CMPS–Feline (2.8 ± 1.6; P = 0.014) and FGS scores (0, range 2–1; P = 0.025) than the social cats (1.3 ± 1.2 and 0, range 0–0, respectively). After surgery, CMPS–Feline (P = 0.011) and FGS (P = 0.009) scores were higher in the independent cats (3, range 3–5 and 2, range 1–3, respectively) than in the social ones (2, range 1–3 and 1, range 0–1, respectively). Within personality groups, there were no differences between study variables measured before and after surgery. The main finding of the present study is that feline personality affects reliability of both the CMPS–Feline and FGS. This should be considered when these scales are used to assess pain in cats, and to adjust the dose of pain killers based on pain assessment.

Introduction

Pain, a multidimensional experience that in humans integrates sensory, affective and cognitive components, is different from mere nociception and is expressed through species-specific behaviours. ¹

In feline patients, these behaviours are often subtle and highly context-dependent. ² As a result, despite the unquestionable benefits of having validated pain assessment tools, challenges are still encountered when veterinarians attempt to evaluate and correctly interpret feline acute pain. ³ Although some limitations are intrinsic to the assessment of non-verbal species, some others are associated with species-specific behavioural expression of pain. ⁴ Domestic cats (Felis catus) are generally reluctant to express their distress, with a tendency to mask illness or injury to a greater degree than socially obligate species. ⁵ In this respect, they are believed to share behavioural similarities with their wild counterparts, namely the solitary predators of the Felis lybica species. However, within the species-specific behaviours, personality dimensions related to sociability and inter-specific communication are influenced by various factors. This results in individuals that act as solitary predators. At the opposite end of the spectrum, certain individuals can manifest gregarious tendencies. ⁶

The common denominator between the perception of pain and an individual’s personality is that both are impacted by the subject’s emotional experience. ⁷ The manifestation of pain, on the other hand, is impacted by the individual’s capacity to express themselves and their motivation to do so. ⁸

Studies in horses showed that certain personality traits correlate closely with behaviour and that extraversion, a distinct personality dimension, has the potential to affect the manifestation of musculoskeletal pain.^9,10 Similarly, dogs that were classified as ‘extraverted’ based on a personality test scored higher on a validated pain scale compared with ‘neurotic’ ones. ¹¹ Personality traits, including personality dimensions such as neuroticism and extraversion, have also been researched in cats, although methodological inconsistencies, such as the use of different assessment tools and heterogeneous study populations, make comparison of results across studies challenging.^{12 –16} Despite this, there is general agreement that sociability or friendliness is a measurable trait in cats, spanning a spectrum from highly social to more independent individuals. ¹⁷

Acute surgical pain in cats is assessed and quantified to allow effective titration of analgesic therapy, thereby improving pain management and enhancing the welfare of feline patients. Both the Feline Grimace Scale (FGS) and the Glasgow Feline Composite Measure Pain Scale (CMPS–Feline) are validated for this purpose and rely on behavioural indicators.^18,19 The FGS is based on cats’ facial expressions; the score (range 0–10) is assigned based on the evaluation of specific facial features, such as ear position, orbital tightening, muzzle tension, and whisker and head positions. ¹⁸ A shorter version of the CMPS–Feline is also available; this behaviour-based assessment scale integrates facial expressions and feline behaviours observed both during interaction and in the absence of interaction. It consists of seven items, each with a different scoring range, yielding a total score of 0–20. ¹⁹ Although these tools are widely used, their reliability may be affected by factors unrelated to pain, such as temperament, personality traits and emotional states.^20,21

Mechanical thresholds (MTs) have been proposed as an alternative semi-quantitative method to evaluate pain in cats and have been used in various clinical scenarios, with conflicting results.^{22 –24} The principle behind MTs is that subjects experiencing pain in a specific area typically show decreased tolerance to mechanical stimulation at that site, resulting in lower MTs compared with either non-painful areas in the same individual (intra-subject comparison) or the same site in other individuals (inter-subject comparison). ²² However, this method has not yet been validated for use in cats.

The primary objective of this study was to determine whether feline personality would affect the scores of both the FGS and CMPS–Feline obtained before and after undergoing elective neutering surgery in a population of healthy cats. The secondary objective was to determine whether MTs, measured with an algometer pre- and postoperatively, would also be affected by feline personality.

In the long term, the aim of the present study was to enhance the understanding of the limitations of currently used feline pain scales and to improve pain assessment and management in cats.

The study hypotheses were that cats identified as ‘social’ would have lower pre- and postoperative FGS and CMPS–Feline scores, and higher pre- and postoperative MTs, than cats classified as ‘independent’.

Materials and methods

Ethical approval

The study was approved by the Ethics and Welfare Committee of the Department of Veterinary Medicine of the University of Cambridge (licence number: CR791; approved on 26 March 2023). Written and/or verbal informed consent was obtained from the owner of each cat before enrolment in the study.

Study design and animals

The study was designed as a prospective clinical trial. Client-owned or fostered healthy male and female cats presented through the primary care service at the Queen’s Veterinary School Hospital (QVSH) of the University of Cambridge for elective neutering were recruited for the study. Exclusion criteria were age less than 3 months, previous diagnosis of any clinical condition, poor tolerance of handling or approach, and aggressive demeanour.

After admission to the hospital, each cat was housed individually in a kennel and allowed an acclimatisation period of 12–14 h.

Procedures and anaesthesia

In the early morning of the day of surgery, the cats were assessed with the Feline-ality personality assessment (see Appendix 1 in the supplementary material) ²⁵ and, based on their score and personality type, assigned to one of two groups (independent and social). For ease of interpretation, cats that scored above 21 were still considered social rather than gregarious.

The personality assessment was performed on one cat at a time, always by the same investigator (AC). The first four items of the personality assessment were evaluated in the cat wards, with the animals in their kennels, while the subsequent seven items were assessed in a novel space, namely the hospital feline consultation room, as per test instructions. ²⁵ The cats were moved from the kennels to the consultation room using a carrier covered with a blanket; transportation time was approximately 2 mins. For item 5, a smartphone timer was used to calculate the interaction time. For item 11, a variety of toys (a plush mouse, a feather wand and a plastic ball the size of a golf ball) were used to incite the cat to play. After the assessment, cats were returned to their kennel.

Approximately 2 h after the personality assessment and before being moved to the operating theatre, a preoperative (baseline) pain assessment was performed on the cats, still in their kennels, by another investigator (CA), with two validated scales: the FGS (score 0–10), always used first, and the CMPS–Feline (score 0–20). Subsequently, MTs (range 0–1000 g) were measured with a mechanical algometer specifically designed for small animals (SMall animal ALGOmeter – SMALGO; Bioseb). The 3 mm tip was mounted on the sensitive probe of the SMALGO and the unit ‘g’ (grams) was selected on the control unit; thereafter, the device was zeroed and the key ‘max’ pressed to enable storage and recording of the maximum force (set at 1000 g) applied during the probe application. After the instrument was set, MTs were measured with the cats in a standing position and the probe applied perpendicularly to the skin at a reference point, namely the left anconeal process of the elbow, with a steadily increasing force until either a behavioural response was observed or the maximal force was reached. Sustained vocalisation, head turning towards the stimulation site, hissing, and attempting to bite, scratch or escape were considered positive behavioural responses.

At the end of the assessments, a standard elective neutering surgery (either ovariectomy via flank laparotomy or orchiectomy) was performed, always by the same surgery team, including one final-year veterinary student on clinical rotation who actively participated in the procedure. A standard anaesthetic and analgesic protocol, including premedication with methadone (0.2 mg/kg IM, Synthadon; AnimalCare), medetomidine (0.01 mg/kg IM, Sedator; Dechra) and ketamine (2 mg/kg IM, Narketan; Vétoquinol), anaesthetic induction with intravenous (IV) propofol (PropoFlo; Zoetis) titrated to effect to achieve endotracheal intubation, and maintenance, via a non-rebreathing system, with inhalational isoflurane (Vetfluran; Virbac) in oxygen, was used consistently for all cats. Postoperative analgesia included meloxicam (0.2 mg/kg SC, Metacam; Boehringer Ingelheim) given at the end of surgery. In female cats, buprenorphine (0.02 mg/kg SC, Buprelieve; Jurox) was administered postoperatively, either 4 h after the administration of premedication with methadone or sooner, in case of postoperative FGS and CMPS–Feline scores higher than 2 and 4, respectively.

Postoperative pain assessments were performed by the same investigator who conducted the preoperative ones (CA), with the same methodology, as soon as the cats were conscious and fully recovered from anaesthesia (able to maintain sternal recumbency and lift up their heads in response to verbal or tactile stimulation).

The two investigators involved in the study (AC and CA) performed their evaluations independently and were blinded to each other’s assessment until the end of data collection.

Statistical methods

To determine the minimum sample size, a validated online sample size calculator (ClinCalc.com) was used. The variables were set with an alpha value of 0.05, study power of 80% and a beta value of 0.2. Based on the study hypothesis, the expected means for preoperative CMPS–Feline and FGS were set at 4 ± 2 and 2 ± 1, respectively, for the cats classified as ‘independent’ and for those classified as ‘social’. The calculation indicated that a minimum of 14 cats per personality type – independent and social – was required to detect a difference in the study outcome variables, for a total sample size of 28 cats. Based on this, and since the cats’ personalities could not be determined during initial enrolment but only after assessment, permission to include up to 40 cats was requested on the ethical application to increase the chances of obtaining groups (‘independent’ and ‘social’) of comparable sizes.

The Shapiro–Wilk test was used to assess data distribution. Depending on the distribution, either the independent samples t-test or the Mann–Whitney U-test was used to compare continuous variables between groups. For comparison of continuous variables between time points within the same group, either the paired t-test or the Wilcoxon signed-rank test was used for normally and non-normally distributed data, respectively. Statistical analyses were performed using commercially available software (SigmaStat 3.5 and SigmaPolt 10; Systat). A P value <0.05 was considered statistically significant.

Results

Data were collected between April and November 2024 and are presented as either mean ± SD or median (interquartile range [IQR]), depending on data distribution.

A total of 29 cats, with a median age of 6 months (IQR 3–30) and mean weight of 2.8 ± 1.1 kg, completed the study. Based on the personality assessment, 14/29 cats were assigned to the ‘independent’ group and 15/29 to the ‘social’ group. Of the 29 cats, 14 were female and 15 were male. There were no differences in body weight, age and sex distribution between the groups.

Surgeries and general anaesthesia were unremarkable, and no perioperative complications were reported. Three cats required additional sedation to allow IV catheter placement, which consisted of IM medetomidine (0.01 mg/kg) and ketamine (1 mg/kg). The mean duration of surgery was 25 ± 6 mins for the orchiectomies and 52 ± 18 mins for the ovariectomies. Postoperative pain assessment occurred a mean of 72 ± 19 mins after tracheal extubation, always within 3 h after methadone administration and, in the female cats, before administration of buprenorphine. There were no differences between groups in the duration of surgery and time elapsed between the end of anaesthesia and pain assessment. All cats were discharged from the hospital on the evening of the surgery.

Preoperatively, the independent cats had statistically significantly higher mean CMPS–Feline scores (2.8 ± 1.6) and median FGS scores (0, IQR 2–1) than the social cats (1.3 ± 1.2 and 0, IQR 0–0, respectively; P = 0.014 for CMPS–Feline and P = 0.025 for FGS). Mean preoperative MTs did not differ between groups (376 ± 190 for the independent cats and 420 ± 142 for the social cats; P = 0.492).

Median postoperative CMPS–Feline and FGS scores were higher in the independent cats (3, IQR 3–5 and 2, IQR 1–3), respectively) than in the social cats (2, IQR 1–3 and 1, IQR 0–1, respectively; P = 0.011 for CMPS–Feline and P = 0.009 for FGS). Mean postoperative MTs did not differ between groups (401 ± 173 for the independent cats and 448 ± 133 for the social cats; P = 0.431).

For intra-group comparisons, no statistically significant differences were found between pre- and postoperative values in either group for CMPS–Feline scores (P = 0.050 for the independent cats and P = 0.131 for the social cats), FGS scores (P = 0.195 for the independent cats and P = 0.533 for the social cats) and MTs (P = 0.808 for the independent cats and P = 1 for the social cats) (Figures 1 and 2).

Figure 1

Box plots representing the CMPS–Feline scores obtained at baseline and postoperatively in 29 cats undergoing elective neutering surgery, for both the ‘independent’ and ‘social’ groups. The placement of the asterisks indicates a statistically significant difference in CMPS–Feline scores between groups (independent/social). CMPS–Feline = Glasgow Feline Composite Measure Pain Scale

Figure 2

Box plots representing the FGS scores obtained at baseline and postoperatively in 29 cats undergoing elective neutering surgery, for both the ‘independent’ and ‘social’ groups. The placement of the asterisks indicates a statistically significant difference in FGS scores between groups (independent/social). FGS = Feline Grimace Scale

Discussion

This study shows that feline personality, determined using the Feline-ality assessment tool, affects the outcome and reliability of two validated pain scales for cats, with subjects classified as independent scoring higher than social cats even preoperatively, presumably in the absence of pain. This suggests that both pain scales may have lower accuracy and reliability when used in cats with an independent personality.

It is acknowledged that cats tend to mask pain-related behaviours, being both a predator and a prey species. ²⁶ This makes the detection and interpretation of pain challenging, even for experienced observers, especially in an unfamiliar environment such as the veterinary setting. ²⁷

An effect of personality on clinical assessment has previously been demonstrated in animal species other than cats. In both horses and dogs, a correlation was found between the degree of extraversion and lameness and pain scores, respectively.^9,10,11,28 Although previous studies assessing equine and canine personalities – mainly focused on extraversion and neuroticism – relied on validated owner questionnaires, the current study assessed feline personality through a direct test based on observation. The Feline-ality Assessment, part of the Meet Your Match Adoption Program of the American Society for the Prevention of Cruelty to Animals (ASPCA), was used for this purpose. ²⁵ Its previous validation in diverse settings supports its suitability for heterogenous feline populations, including not only client-owned cats but also foster cats and stray cats cared for by rescue organisations.^16,29,30 Alternative personality assessment tools could also have been used. For example, the Feline Five assumes personality traits similar to those examined in equine and canine studies and is based on the same concept. ¹²

There are controversies regarding the types of personalities existing in cats. One study suggested that domestic cats have six personality dimensions, whereas another indicated that they have seven.^14,15 Gartner and Weiss ¹³ found that the most consistent personality dimensions across the feline population are ‘sociable’, ‘curious’ and ‘dominant’. Other aspects of feline personality that have previously been investigated include the response to acute confinement, based on which two distinct coping styles, ‘reactive’ and ‘proactive’, have been identified. These two styles are influenced by personality dimensions, with the reactive style associated with more timid individuals, while the proactive style is characteristic of those that are playful and inquisitive. Although Feline-ality also includes assessment of valiance, this aspect was not evaluated in the current study, as the focus was specifically placed on the personality dimension of sociability. ²⁹

In the literature, terminology is often used inconsistently, with terms such as ‘personality’, ‘behavioural traits’, ‘syndromes’, ‘temperament’ and even ‘demeanour’ being used interchangeably; however, although behaviour and demeanour are deterministic and can vary based on specific situations, personality represents a more constant and unique characteristic of the individual.^21,31 A previous study investigated the impact of demeanour on pain assessment with two validated pain scales, namely the CMPS–Feline and the UNESP-Botucatu Multidimensional Composite Pain Scale. ²¹ Other authors have reported the use of a demeanour scoring system, assigning cats to 1/5 behavioural categories: friendly and confident, friendly and shy, withdrawn and protective, withdrawn and aggressive, and overtly aggressive. ³² The study also found that cats required 2 days to acclimatise to the hospital environment. In the present study, cats were acclimatised for 12–14 h before assessment. According to the Feline-ality manual and training guide, an 18-h acclimatisation period is considered sufficient to provide a reliable personality prediction. ²⁵ Although this remains a recognised limitation of our study methodology, an acclimatisation period longer than 14 h was not feasible owing to constraints related to owner availability and the need to minimise disruption to hospital workflows. Timing and environment suitability were other reasons for not waiting 18 h to assess the cats. The cats were typically admitted between 4 and 6 pm on the day before surgery (depending on client availability) and assessments were conducted between 5 and 7 am the next morning. Early morning was considered the best time for conducting assessments because of reduced staff presence, noise and interactions with the cats – minimising potential distractions that could have interfered with the personality assessment.

Another potential limitation is the relatively young age of some of the cats included in the study. The age cutoff for inclusion was set at 3 months, as this is a common age for feline surgical neutering in the UK. In addition, at the institution where this study was conducted, elective neutering of adult cats is uncommon. For adoption and matching programmes, it is generally recommended not to assess feline personality in cats younger than 9 months. However, there is currently no consensus on when feline personality is fully developed or at what age it stabilises.^{33 –37} A previous study investigating feline friendliness towards humans found this trait to be consistent between 3 and 8 months of age, suggesting that setting a higher cutoff may not have significantly affected the findings of the present study. ³⁸ Based on these considerations, the findings of this study can be confidently applied to young cats, although greater caution is advised when applying them to adult and senior cats.

In addition to age, another challenge encountered during the study design was the standardisation of the study population in terms of the intensity of surgical stimulation, as well as intra- and postoperative nociception and pain. Ovariectomy involves a flank laparotomy and is therefore considered more invasive than orchiectomy; however, as with other demographic variables, the sex of the cats was evenly distributed between the personality groups. To minimise the influence of potential confounding factors, the timing of assessments was standardised as much as possible, and all cats were evaluated under similar analgesic conditions, within the effective duration of preoperative methadone and after administration of meloxicam. Although postoperative assessments may have been influenced by differences in pain perception and variability in surgical stimulation because of the teaching environment, it is reasonable to assume that the study population was adequately standardised for preoperative pain evaluations, as these were performed in healthy, pain-free cats under similar conditions and in the same setting.

A further limitation of this study was the absence of multiple observers for each type of assessment. The reason for this was that both personality and pain evaluations involved direct handling and interaction with the cats. Including additional observers could have resulted in reduced feline compliance and potentially increased stress levels during hospitalisation and the perioperative period, which would have not only affected the study outcomes but also raised ethical concerns.

MTs were affected neither by personality nor by timing of the assessment. This finding was unexpected; however, it is worth considering that MT measurement is not a validated method for pain assessment in cats. Ideally, MTs should have been obtained in proximity to the surgical wound; however, the inclusion of both male and female cats would have introduced an additional variable represented by the anatomical area were the measurements were taken. For this reason, a reference point was selected instead. Regarding the suitability of the assessment tools for this study, it should be noted that although the Feline-ality Assessment has been validated in various contexts, it was originally developed as part of an adoption matching programme, rather than for evaluating feline personality in a clinical setting.

The absence of differences between pre- and postoperative pain scores in both groups may reflect either a lack of sensitivity of the pain scales in detecting postoperative pain or, more likely, the result of appropriately managed perioperative analgesia used in the study cats.

Conclusions

This study demonstrates that personality type influences the reliability of behaviour-based pain assessment scales in young cats, which may result in overestimation of pain, and potentially unnecessary administration of analgesics, in independent cats.

Supplemental Material

Appendix 1

Description of the criteria for the Feline-ality assessment and instruction how to perform it.