Use of mechanical thresholds in a model of feline clinical acute pain and their correlation with the Glasgow Feline Composite Measure Pain Scale scores

Abstract

Objectives

The aim of this study was to evaluate the use of mechanical thresholds (MT), measured with the SMALGO (Small Animal ALGOmeter), and to determine whether there was a correlation between MT and Glasgow Feline Composite Measure Pain Scale (CMPS–Feline) scores in cats undergoing ovariohysterectomy.

Methods

Client-owned cats undergoing flank ovariohysterectomy were recruited. Pain scores for the pre- and postoperative periods were obtained using the CMPS-Feline in each cat by two independent investigators (A and B). Following CMPS–Feline scoring, MT were measured with the SMALGO, in the surgical area, pre- and postoperatively, only by investigator A. Each cat served as its own control for the comparison of pre- and postoperative variables. Reliability statistics were used to assess the level of inter-observer agreement (A vs B) with respect to pre- and postoperative CMPS–Feline scores, while Spearman’s correlation statistics were used to analyse the relationship between MT and CMPS–Feline scores.

Results

Twenty-nine cats completed the study. Preoperative MT (340 g [range 108–691]) were significantly higher than postoperatively (233 g [range 19–549]; P = 0.001). CMPS–Feline scores were not found to differ significantly between the preoperative period (2 [range 0–7] for investigator A and 3.2 ± 2.3 for investigator B) and postoperative period (2 [range 0–10] for investigator A and 3 [range 0–8] for investigator B) for either investigator. Reliability statistics revealed that the level of inter-observer agreement with respect to CMPS–Feline was fair for the preoperative assessments but poor for the postoperative evaluations. There was no correlation between MT and CMPS-Feline scores.

Conclusions and relevance

Although there was no correlation with CMPS–Feline scores performed at the same timepoint, MT increased postoperatively vs baseline. Assuming that, despite analgesia, susceptibility of the surgical area to mechanical stimulation would increase after surgery, this finding suggests that MT might be useful to assess feline surgical pain. The poor level of inter-observer agreement with respect to postoperative CMPS–Feline scores highlights the potential limitations of this scale.

Introduction

Pain assessment is notoriously challenging in cats. Underestimation of feline pain is perceived as a relevant clinical problem as it is likely to result in insufficient provision of analgesia, and therefore suboptimal patient care. ¹

As a result, various methods aimed at evaluating and measuring pain have been developed for cats and investigated over time. ¹ Scales such as the Glasgow Feline Composite Measure Pain Scale (CMPS–Feline), the UNESP–Botucatu Multidimensional Composite Pain Scale (UNESP–Botucatu MCPS) and the Colorado State University Feline Acute Pain Scale (CSU-FAPS) have been developed and validated for acute pain specifically.^2–7 While they are commonly used and are seemingly simple to perform, there is an inherent degree of subjectivity when using them. One of their limitations is that they rely on behavioural indicators, namely posture, facial expressions and reaction to stroking/palpation that can be affected by fear and anxiety, especially when the cats are assessed in an unfamiliar environment such as the veterinary hospital. In recent years, the use of the CMPS–Feline has increased in small animal practice in the UK, and is now commonly regarded as a useful and reliable tool to detect and address inadequate postoperative analgesia.^8,9

Nociceptive threshold testing techniques are methods of assessing nociceptive function and reaction after stimulation with sensory inputs like pressure, heat or cold. The value at which point the animal is deemed to react to the sensory input is taken and used for statistical analysis. ¹⁰ Nociceptive threshold testing techniques have been investigated in cats as a complementary tool for pain assessment, attempting to reduce the level of intra- and inter-assessor variability and therefore provide a higher degree of objectivity.^10–14 Sensory mechanical thresholds (MT) are obtained by applying and measuring either a force or a pressure on a targeted area of the body of the animal until a specific, predefined behavioural response is observed. Various mechanical ALGOmeter have been used in feline experimental and clinical models.^11–15

Among these, the Small Animal ALGOmeter (SMALGO; Bioseb) was originally developed for use in laboratory rodents. In both healthy cats and cats with chronic pain, the SMALGO has been shown to give repeatable results from users with different backgrounds and degree of expertise in pain assessment. ¹¹ When used to assess cats with chronic gingivostomatitis, the MT measured with this device did not correlate with the severity of the disease. ¹¹ To our knowledge, the SMALGO has yet to be investigated in a model of feline acute pain.

Ovariohysterectomy is one of the most common surgical procedures performed routinely in cats in the UK. ¹⁶ As most cats are presumably non-painful before surgical neutering is performed, ovariohysterectomy may be regarded as a suitable model of acute pain in otherwise healthy cats. Reliable and objective assessment of acute pain would result in prompt and targeted administration of rescue analgesics in cats undergoing spay surgery. It is our opinion that this could improve the quality of care and welfare of a large number of cats worldwide.

The primary objective of this study was to evaluate the use of the SMALGO as a method of quantifying pain in cats undergoing routine spay, and to determine whether MT would change after surgery. The secondary objective was to assess for correlation between the MT measured with the SMALGO and the scores obtained with the CMPS–Feline, both pre- and postoperatively.

It was hypothesised that MT would decrease after surgery while the CMPS–Feline scores would increase, and that there would be an inverse correlation between the two variables both pre- and postoperatively.

Materials and methods

Ethical approval

The study was conducted with permission from the Clinical Research Ethical Review Board of the Royal Veterinary College (licence number: URN 2019 1909–3; approval date 25 September 2019). Written informed consent was obtained from the owner of each cat before enrolment in this study.

Study design

The study was designed as a prospective clinical trial. Client-owned female cats presenting to a primary care practice (Medivet 24 Hour Hendon, London, UK) for flank ovariohysterectomy were recruited for the study. Exclusion criteria were cats <6 months or >3 years of age, pregnancy, previous diagnosis of any neurological condition (as they may alter the physiology of pain transmission and perception), aggressive behaviour or poor tolerance to handling, and any analgesic treatment that may have increased the MT and/or alter behavioural pain scores.

The study variables were pre- and postoperative CMPS–Feline scores, and pre- and postoperative MT. All data were collected by two experienced and trained registered veterinary nurses (investigator A and investigator B). After admission to the hospital, each cat was moved to a kennel and allowed an acclimatisation period of at least 15 mins. Investigators A and B both independently, in a non-randomised order, used the CMPS–Feline to derive a baseline preoperative score, which was then recorded. Each investigator’s assessment of the cat and the score they derived was blinded from the other. Thereafter, investigator A, who had been previously trained by the authors (DN and CA), measured the MT with the SMALGO in each cat. MT were measured as follows: the sensitive probe of the SMALGO was equipped with the 3 mm tip and the unit ‘g’ (grams) selected. Thereafter, the control unit was zeroed and the key ‘max’ pressed, to enable storage and recording of the maximum force value applied during the probe application. With the cats in standing position, the SMALGO probe was then applied perpendicularly to the skin of the left flank, 1 cm caudal to the middle of the surgical incision site (Figure 1), with a steadily increasing force, until either a behavioural response was observed or the maximal force was reached. Vocalisation, head turning towards the stimulation site, back muscle contraction, hissing and attempting to bite, scratch or escape were considered positive behavioural responses. The maximal cutoff force was set at 400 g, based on previous studies. ¹¹ The force measured was recorded as MT.

Figure 1

Diagram of a cat demonstrating the positioning of the Small Animal ALGOmeter (SMALGO) probe for mechanical threshold testing in this study. The application site was 1 cm caudal to the middle of the surgical incision area, on the left flank

After the preoperative assessments, cats were premedicated with intramuscular medetomidine (Sedator 0.01 mg/kg; Dechra) and buprenorphine (Vetergesic 0.03 mg/kg; Ceva Animal Health) before the placement of an intravenous (IV) catheter once the cat was deemed to be adequately sedated for this to be performed. Anaesthesia was then induced with IV propofol (PropoFlo; Zoetis), titrated to effect to allow endotracheal intubation. Anaesthesia was then maintained with inhalational isoflurane (IsoFlo; Zoetis) in 100% oxygen, delivered via an Ayre’s T-piece non-rebreathing breathing system connected to an appropriate size endotracheal tube. Ovariohysterectomy was then performed by a veterinary surgeon (JD) via a left-sided flank approach. Subcutaneous meloxicam (Metacam 0.2 mg/kg; Boehringer Ingelheim Animal Health) was given postoperatively to each cat, at the end of the anaesthetic after endotracheal extubation.

Postoperative assessments were performed when the cats were conscious and responsive following general anaesthesia, shortly before they were discharged from the hospital. Postoperative assessments were performed in the same manner as preoperative: CMPS–Feline by investigators A and B, independently, and MT measured by investigator A. Intervention level for re-evaluation of the postoperative analgesic plan by a veterinary surgeon was set at score of 5/20 or higher, as assessed by both investigators, of the CMPS–Feline. ² Cats with a score of ⩾5 received rescue analgesia consisting of methadone (Comfortan 0.2 mg/kg IV; Dechra) every 4 h as required based on regular pain assessment.

Statistical methods

The sample size was calculated separately for the two different aims of the study. In order to detect a difference between pre- and postoperative MT with either a paired t-test or the non-parametric Wilcoxon signed rank test, the expected difference in means (preoperative – postoperative values) was set as 100 g, with an expected SD of 100 g, desired power of 0.9 and alpha (α) of 0.05. This calculation indicated 23 as the minimum number of subjects to be included in the study. A second calculation was performed with the purpose of investigating the relationship between the variables MT and CMPS–Feline score. Variables were set as follows: α value: 0.05; power: 0.8; beta (β) value: 0.2; type of test: two-sided test; standard normal deviation for α (Zα): 1.960; standard normal deviation for β (Zβ): 0.842; expected correlation coefficient (r): –0.5. ¹⁷ This resulted in a minimum number of subjects of 29. Data distribution was analysed with both the Kolmogorov–Smirnov and the Shapiro–Wilk tests. The non-parametric Wilcoxon signed rank test was used for paired comparison of pre- and postoperative MT.

Reliability statistics were used to analyse the level of inter-observer agreement between investigators A and B, which was scored as follows: intraclass correlation coefficient (ICC) <0.40 = poor agreement; ICC 0.40–0.59 = fair agreement; ICC 0.60–0.74 = good agreement; ICC 0.75–1 = excellent agreement. ¹⁸

Spearman’s rank order correlation statistics were used to determine the relationship between MT, following logarithmic (base-10) transformation, and demographic variables (age, body weight and body condition score [BCS]), and between MT, following logarithmic (base-10) transformation, and CMPS–Feline scores.

Commercially available statistical software (SigmaStat 3.5 [Systat Software] and SPSS Statistics 26 [IBM]), as well as a validated online calculator from the Clinical and Translational Science Institute of the University of California (https://sample-size.net/correlation-sample-size/) were used. ¹⁹ P values <0.05 were considered to be statistically significant.

Results

Data are presented as either mean ± SD, or median (range), depending on the data distribution.

Data were collected between November 2019 and June 2020. Twenty-nine cats aged 10 months (range 6–26) were included in the study; body weight and BCS were 3.1 ± 0.5 kg and 4/9 (range 3–5), respectively. Surgery and anaesthesia were unremarkable for all cats. Three cats had a postoperative CMPS–Feline score of ⩾ 5 assigned by both investigators, and so remained hospitalised for the administration of rescue analgesia. They remained hospitalised for 12–18 h and were discharged the day after surgery. The cats who received rescue analgesia and remained hospitalised were included in the data; however, only the CMPS–Feline scores and MT from the immediate postoperative period were analysed, alongside the preoperative assessments. Any pain assessments after the initial postoperative period were not included in the data.

Preoperative MT (340 g [range 108–691]) were significantly higher than postoperative MT (233 g [range 19–549]; P = 0.001) (Figure 2). There were no statistically significant differences between the CMPS–Feline scores obtained by investigators A and B, neither preoperatively (2 [range 0–7] and 3.2 ± 2.3, respectively) nor postoperatively (2 [range 0–10] and 3 [range 0–8], respectively) (Figure 3). Five cats, assigned a preoperative CMPS-Feline score >5, were also assessed by the veterinarian in charge performing the surgery, who decided on the basis of accurate physical examination and knowledge of the patient’s history to exclude potential causes of pain.

Figure 2

Box plots represent the mechanical thresholds (MT), measured with the Small Animal ALGOmeter (SMALGO) pre- and postoperatively by two independent investigators (A and B), in 29 female cats undergoing spay surgery. The upper and lower quartiles (interquartile range box) represent the data greater (25%) and lesser (25%) than the median, respectively, accounting for 50% of the total data. The whiskers represent the ranges for the bottom 25% and the top 25% of the data values. The dots represent the outliers

Figure 3

Box plots represent the Glasgow Feline Composite Measure Pain Scale (CMPS–Feline) scores obtained preoperatively (1) and postoperatively (2), by two independent investigators (A and B), from 29 female cats undergoing spay surgery. The upper and lower quartiles (interquartile range box) represent the data greater (25%) and lesser (25%) than the median, respectively, accounting for 50% of the total data. The whiskers represent the ranges for the bottom 25% and the top 25% of the data values. The dots represent the outliers

Reliability statistics revealed that the level of inter-observer agreement between investigators A and B for CMPS−Feline was fair for the preoperative assessments (ICC 0.545, 95% confidence interval [CI] 0.03−0.79) but poor for the postoperative evaluations (ICC 0.312, 95% CI −0.47 to 0.68).

No statistically significant correlations were detected between preoperative MT and CMPS−Feline scores obtained by either investigator (A: correlation coefficient [CC] 0.038 [P = 0.843]; B: CC −0.08 [P = 0.663]), between postoperative MT and CMPS−Feline scores obtained by either investigator (A: CC 0.02 [P = 0.920]; B: CC 0.25 [P = 0.188]) (Figure 4), or between MT and demographic variables, namely age (CC −0.05; P = 0.69), body weight (CC −0.10; P = 0.45) and BCS (CC −0.23; P = 0.07).

Figure 4

Scatterplot representing the correlations between the Glasgow Feline Composite Measure Pain Scale (CMPS−Feline) scores and mechanical thresholds (MT) measured with an ALGOmeter. Assessments were performed preoperatively (preop) and postoperatively (postop), by two independent investigators (A and B), in 29 female cats undergoing spay surgery

Discussion

This study showed a decrease in MT postoperatively, but no correlation was evident with CMPS–Feline scores in cats undergoing routine ovariohysterectomy. Furthermore, the level of agreement between two experienced nurses with respect to the postoperative CMPS-Feline score was unexpectedly poor.

The changes in MT in the study cats were similar to that of some previous studies. In one study conducted in male cats undergoing castration, MT decreased for up to 2 h after surgery and such a decrease was greater in the control group, which did not receive pethidine. ¹³ Another study comparing the use of methadone to buprenorphine and butorphanol in both male and female cats undergoing surgical neutering found that MT decreased postoperatively in female cats, when compared to preoperative baseline in both the buprenorphine and butorphanol groups. ²⁰

Although adequate analgesia is expected to increase MT, it is reasonable to assume that cats who underwent surgery would have increased sensitivity to touch and pressure applied to the surgical area, and therefore their postoperative MT would still be lower than preoperative values, even when pain therapy has been administered. The decrease in MT after surgery was therefore interpreted as a sign of pain in the study cats, although decreased tolerance to mechanical stimulation, which has been described in cats after repeated testing, cannot be completely excluded. ¹² Concern for decreased tolerance after repetition of testing was the reason why, when the study was designed, it was decided that only one of the investigators would measure MT. One study assessing the reliability of the SMALGO in healthy cats, found the effect of repetition was only seen after more than two testing sessions within time intervals shorter than 45 mins, suggesting measurements taken more than 45 mins apart should lead to minimal increases in sensitisation or decreased tolerance to handling. ¹² However, in a previous study that used mechanical stimulation to assess analgesia after neutering, three consecutive preoperative MT values were obtained from each cat, as well as several postoperative MT measurements performed at hourly intervals. ²⁰ In that study, no degree of tolerance to testing was found, which suggests that the approach taken in the current study to minimise the effects of repetition of measurements may have been overly cautious.

As MT seem to correlate with the severity of both acute and chronic pain in various animal species, we hypothesised that the study cats would show some degree of correlation with the scores of the CMPS–Feline, a recognised and validated scale to assess acute pain in cats.^21,22 However, this hypothesis was disproved as CMPS–Feline scores were not found to correlate with MT values. This could be due to lack of validity of either MT or CMPS–Feline scores, or both, as tools for the assessment of feline pain after surgery. However, the unexpectedly poor level of agreement between the two investigators with respect to postoperative CMPS–Feline scores raises some concerns regarding the repeatability and objectivity of assessments when this scale is used.

One possible explanation for this finding is that there is some degree of inherent subjectivity within the CMPS–Feline. Although the investigators involved were both experienced registered veterinary nurses, who are familiar with the use of the CMPS–Feline, it is still possible that the subjectivity of the scale led to results inconsistent with the degree of pain the cats were experiencing. The degree of subjectivity seen with CMPS–Feline may partly be due to how literally the wording in the scale is taken by the examiner. To corroborate this theory, one of the study cats was assigned preoperative CMPS-Feline scores of 6 and 0 by investigators A and B, respectively. The cat was vocalising very often but intermittently, a behaviour that investigator A scored by literally compiling the scale; she therefore assigned a score of 1 and 2 to questions 1 and 6 of the CMPS–Feline. Investigator B, who presumed the cat was not in pain, despite having observed the cat vocalising, chose to answer the same questions based on the time in which the cat was intermittently silent, and assigned therefore a score of 0 to both questions 1 and 6. As a result, the CMPS–Feline scores assigned independently by the two investigators for this same pain-free cat were very different. This may have been the case for other cats with preoperative CMPS–Feline scores >6, which should be not possible for presumably non-painful cats. The CMPS–Feline does not allow for differentiation between behaviours indicative of fear or anxiety, and pain. Behaviours such as cowering at the back of the kennel or growling at the examiner, which can be caused by either stress/anxiety or pain, add points to the overall scale, which may reach moderate scores, even in a patient who is non-painful but is uncomfortable in the hospital setting. As an attempt to avoid the inclusion of anxious and fearful cats, aggressive behaviour and poor tolerance to handling were listed among the exclusion criteria. Nevertheless, it is possible that some study cats who were deemed non-fearful on admission to the practice developed anxiety later, after separation from their owner – a drawback that could have affected CMPS–Feline scoring. One way in which this could be prevented in future studies would be the use of habituation periods >15 mins for the cats to settle into their new environment before assessment. However, this may be difficult to achieve in busy clinical practices. For this reason, the limitations of the CMPS–Feline when used in fearful and anxious cats likely represent issues commonly seen in clinical practice.

In order to draw more solid conclusions regarding the use of validated pain scales and their correlation with MT in a model of acute surgical pain, one possible strategy could be to increase the number of assessors, and possibly include other validated scales, such as UNESP–Botucatu, in the study protocol. ⁴ Nevertheless, increasing assessor numbers would have been challenging in this study while maintaining the intense workflow of a busy practice. Moreover, for the reasons mentioned earlier, increasing the number of MT measurements may not improve reliability.

Neutering was chosen as a model of acute pain for this study owing to the fact it is a very common procedure female cats undergo in the UK; however, there are some limitations that come with this choice. One of these is the practicality of the use of the SMALGO on an area of tissue that is large and distensible such as the flank. This could potentially lead to distribution of force over a greater surface than intended as increasing pressure is applied. This limitation was attempted to be mitigated by the choice of the pin-shaped 3 mm tip and by being applied in a perpendicular fashion adjacent to the incision site. Moreover, the type of pain caused by neutering is a complex combination of factors and activation of different receptor types. Mechanical threshold measurement relies on pressure being applied externally which may be assessing somatic pain more than any visceral component. When this is accounted for, it may be unrealistic to expect one testing modality such as MT to assess pain which is inherently multifactorial, and a more comprehensive pain assessment involving multiple stimuli is likely required. ²³ Despite this, MT has been used to assess analgesic efficacy after the neutering of female cats in several previous studies.^20,24 A third limitation associated with neutering model in clinical patients is that modern multimodal analgesic protocols are usually effective, making it more difficult to find correlations where they may actually exist. ²⁵

During this study, the SMALGO was found to be simple to use by the investigator and did not become broken or damaged at any point during its use or storage. These factors are to the SMALGO’s benefit in being used in a busy clinical practice setting, where robust nature and intuitive design are key features to ensure integration into pain assessment.

Conclusions

The SMALGO may have potential applications as a tool for assessment of acute postoperative pain in cats. However, the lack of correlation between MT and CMPS–Feline scores in this study, as well as the poor inter-observer agreement with respect to postoperative CMPS–Feline scores, seem to suggest that more prospective studies are needed to improve the methods of pain assessment in cats, and to clarify the role of mechanical testing in measurement of feline acute pain.