Evaluation of music therapy to reduce stress in hospitalized cats

Abstract

Objectives

This study aimed to evaluate the use of two different types of music – cat-specific music and classical music – compared with no music, to reduce stress in cats during hospitalization.

Methods

Thirty-five hospitalized cats were randomly divided into three groups and each group received a different stimulus – cat-specific music, classical music or no music (control) – throughout their hospitalization. Respiratory rate, salivary cortisol and social interaction were documented. A blinded researcher performed the Cat Stress Score (CSS) during the video analysis of recordings at five specific times over 31 h of hospitalization.

Results

There was no difference in the mean CSS between cats listening to cat-specific music, classical music and control throughout the five evaluations. Cat-specific music had a higher percentage of positive social interactions than the other groups on the first evaluation (P <0.05). The average respiratory rate was significantly lower in the classical music group vs control on the fourth evaluation (P <0.05). Although statistically insignificant, the average respiratory rate decreased only in the classical music group during the five evaluations. Cortisol quantification did not seem to follow the CSS results. However, owing to the low and unrepresentative number of samples, it was not possible to perform statistical analysis on these results or a group sample comparison.

Conclusions and relevance

Both cat-specific music and classical music seem to have some benefit to hospitalized cats. The salivary cortisol analysis was not adequate nor useful to measure stress in hospitalized cats in our study.

Introduction

The positive psychological and physiological effects of music have been well reported in humans.^1–7 Changes in brain physiology, cognition, chemistry and morphology induced by music have been studied in animal models and have provided evidence that music can affect animals the same way it affects humans. The potential benefits of music in animal wellbeing include environmental enrichment, stress reduction and behavioral modulation. ⁸ Likewise, music can be beneficial in masking potentially sudden artificial and stressful noises, as well as breaking the silence in a monotonous environment. ⁹ The goal of environmental enrichment is to help confined animals deal with their daily environmental challenges, such as the feeling of lack of control or unpredictability, encouraging normal behaviors for the species, and reducing abnormal or stereotyped behaviors. ¹⁰

Music’s beneficial effects have been reported in other animals, including: reduction in stereotyped behaviors in elephants; ¹¹ reduction in chimpanzee aggressiveness and an increase in their rest time; ¹² increase in the milk production of dairy cows and an easier approach to the milking parlor when music was playing;^13,14 and reduction in the barking frequency of sheltered dogs and, when classical music was playing, more time resting and less time standing. ¹⁵ Bowman et al also studied dogs, and they found no significant reduction in barking; however, barking increased after ceasing the sound stimulus. ¹⁶

Several environmental aspects can affect the well-being of cats when confined in a hospital environment. ¹⁷ Chronic stress is undesirable not only for humanitarian reasons, but also because it can negatively affect the immune system,^18,19 which may develop or reactivate infection. ²⁰ Besides, cats under stress may reduce their grooming and eating, and increase their vigilance, which may compromise their recovery and discharge.^21,22

Increased noise levels, such as staff conversation, heating and cooling systems, alarms and animal noise, combined with novel environments, restraint for treatments and medical procedures, may culminate in harmful stress levels and increased morbidity in hospitalized veterinary patients.^18–24

The use of music should be carried out with the awareness of the perception of the species-specific range of sound sensitivity and how it may vary between species. ⁹ Snowdon et al suggested that music would be more effective if it were appropriate for the sensory and communication systems of the species under study. ²⁵ Based on this, the authors composed songs for cats and reported that cats showed greater interest in this music, compared with human classical music. These songs contain frequencies similar to cat vocal ranges and are composed to create an affiliative effect on cats using pulses related to purring (1380 bpm) and suckling (250 bpm). ²⁵ Recently, another study demonstrated the advantage of cat-specific music over classical music or silence in a veterinary consultation. ²⁶ A study in dogs compared classical music to music composed for dogs, and found that classical music was more effective in mitigating stress. ²⁷

Our study aimed to evaluate whether music with different auditory stimuli – classical music and cat-specific music – compared with no music – can reduce stress signs in cats in the hospital environment. Stress indicators include the Cat Stress Score (CSS), ²⁸ social interaction (SIC), respiratory rate (RR) and salivary cortisol. To the best of our knowledge, this is the first study to evaluate the effect of music in a hospital environment with non-experimental cats.

Materials and methods

Animals

The present study was approved by the Animal Ethics Committee (CEUA/UFRGS, approval 35922). The owners of each animal signed an informed consent form to participate in this study.

Inclusion criteria were cats with a minimum hospitalization of 31 h (cats arrived at least 1 h before the first evaluation at 11:30 am and stayed until the last evaluation on the following day at 5:30 pm), no severe/decompensated respiratory dysfunction, the absence of motor injury that would prevent them from assuming natural positions for the species, normal state of consciousness and no evidence of therapy with corticosteroids.

Study design

This was a prospective, randomized and partially blinded study. Each cat underwent one of the three different auditory tests (cat-specific music, classical music or control). We generated a random distribution plan using an open-access website (http://www.randomization.com). The methods are summarized in Figure 1. The website chose auditory tests to start from 7:30 am until 11:30 am and from 1:30 pm until 5:30 pm in the hospitalization room at the Feline Medicine Service in the Universidade Federal do Rio Grande do Sul’s Veterinary Teaching Hospital. The time intervals were chosen because they contained the greatest flow of people in the day, which concentrates the handling time of the cats. The average number of people present in the hospitalization room during the study period was five (minimum one and maximum 10 people). The Feline Medicine Service (MedFel) is certified with a gold standard status by the American Association of Feline Practitioners’ (AAFP) Cat Friendly Practice Program. The cat’s area is physically separated from the dog’s area and contains nine cages arranged in three vertical rows and three horizontal rows. Each cage is 90 cm width × 60 cm height × 60 cm depth, with a shelf and a hiding place (Figure 2).

Figure 1.

Summarized methods of the five assessments (A1–A5). CSS = Cat Stress Score; RR = respiratory rate

Figure 2.

Cages used in the study, located in the cat hospitalization room in the Feline Medicine Service of the Veterinary Teaching Hospital, UFRGS

Cats that met the inclusion criteria were followed for at least 31 h. As auditory stimuli were placed in the entire environment, all owners who had their cats hospitalized during this time had to agree with the presence of music, even if their cats did not participate in the study. We assessed the cats at five times (A1–A5). A1 was measured 1 h after the cat arrived, after the cat was medicated if needed and manipulated. Each assessment was video-recorded and had the audio removed, allowing a blinded observer to score each cat’s behavioral responses using the CSS. The CSS describes seven possible stress levels of a cat based upon posture and behavioral elements. The scores ranges from ‘fully relaxed’ (score 1) to ‘extremely stressed’ (score 7) (see the supplementary material). ²⁸ We collected saliva during the second and fourth assessments, and evaluated RR and SIC during all assessments. SIC was positive when the cat showed positive emotions, like rubbing when the researcher approached and touched the cat’s chin, and negative when the cat had negative emotions when approached or touched, such as walking away, crouching, showing body tension, facial expression of fear and/or ears flattened or turned back, hissing, dilated pupils and other signs (Table 1). SIC was neutral when the cat did not react when approached or touched. All SIC evaluations were performed by the same observer and in the same way. There were three times in which the SIC scores were carried out: observation in front of the cat’s cage; a slow approximation to open the cage; and hand approximation to try to touch the cat. The researcher tried not to make eye contact with cat, to avoid appearing confrontational and did not speak during the test.

Table 1.

Social interaction score created by the authors to evaluate cats’ responses during the study

Score	Behavior description
Positive	When the researcher approaches, the cat’s behavior or facial expression does not change The cat walks towards the researcher in a friendly manner or stays still When the researcher’s hand touches the cat, it responds to petting, pushing the petted area closer The cat may be shy and motionless, but when the researcher’s hand touches the cat, it responds in a positive manner by pushing the petted area closer
Neutral	When the researcher approaches, the cat’s behavior or facial expression does not change When the researcher’s hand touches the cat, the cat’s behavior or facial expression does not change
Negative	When the researcher approaches, the cat looks scared and attempts to escape or it freezes. If the researcher’s hand approaches the cat, agonistic behaviors worsen

Music

For classical music, we selected a playlist generated on Spotify, with pieces of music from previous studies.^25,29,30 The classical music included only instrumental pieces, and were based on the piano. For cat-specific music, we selected the playlist performed and produced by David Teie, available on Spotify. Music was played via wireless Bluetooth speakers (JBL) from 7:30 am to 11:30 am and from 1:30 pm to 5:30 pm. Speaker volume was set manually each day to around 60 dB and was kept the same throughout the study.

Salivary cortisol

We collected saliva with a specific swab (Salivette; Sarstedt) placing the swab inside the cats’ oral cavity, on the labial commissure region, and leaving it there until the cotton was visibly moist. The analysis required a minimum volume of 25 µl. We immediately froze samples at –20°C and, on the day of analysis, thawed the samples to room temperature and centrifuged at 1500 × g for 15 mins. Then we analyzed cortisol using a commercial immunoassay kit (Salivary Cortisol ELISA Kit; Salimetrics) at the laboratory.

Statistical analysis

We performed descriptive analyses of all variables. Regarding categorical variables, we used simple frequency tables, with absolute and relative frequency, as well as cross tables, when convenient. For quantitative variables, we calculated minimum, maximum, mean and SD. As for the scores, we calculated the means of the evaluations. The tests performed were a Wilcoxon non-parametric test, to verify whether there was a difference between the first and the last score; a one-way ANOVA test, to assess the difference in mean CSS and mean RR between the groups; Tukey’s test, to verify which groups differed in resonance frequency; Student’s t-test, to compare females and males with the average score; and a χ ² test, to analyze the association of SIC with each group. The level of significance adopted was 5%.

Results

Thirty-eight cats were initially included in the study, but three were excluded because their clinical signs worsened, meeting the exclusion criteria. Thirty-five cats completed at least four of the assessments required. Assessments conducted during the postanesthetic period or sedation were excluded. The average room temperature during the study was 24.1 ± 1.9°C.

There were 22 (63%) males and 13 (37%) females, with a mean age of 5.8 ± 4.1 years. Other demographic data are shown in Table 2. There were no significant differences in age (P = 0.6) or sex (P = 0.4) among groups.

Table 2.

Demographic data of cats divided according to music group

Variables	Cat-specific (n = 11)	Classical (n = 12)	Control (n = 12)
Sex
Male	7	9	6
Female	4	3	6
Reason for hospitalization
FLUTD	2	8	4
CKD	2	0	1
Liver disease	0	1	2
Skin wound	0	1	1
Bone fracture	2	0	0
Small bowel biopsy	0	1	1
Pyometra	0	1	1
SCC	2	0	0
FeLV	1	0	1
Diaphragmatic hernia*	1	0	0
Aggression investigation	1	0	0
Lameness	0	0	1
Mean ± SD age (years)	6.1 ± 4.6	4.5 ± 2.2	6.7 ± 5.0

^*Stable and with a respiratory rate within the reference interval

FLUTD = feline lower urinary tract disease; CKD = chronic kidney disease; SCC = squamous cell carcinoma; FeLV = feline leukemia virus

The mean CSS of the five assessments in the cat-specific music group was 2.5 ± 0.1, 2.4 ± 0.2 in the classical music group and 2.7 ± 0.1 in the control group. There was no difference between the mean CSS in the groups (P = 0.44). No significant differences were found in CSS between males and females (P = 0.5) and between cats older and younger than 8 years of age (P = 0.4).

In the first assessment, a significant association was found in SIC evaluation: the χ ² test showed that the cat-specific music group had more positive interactions than the other groups (P = 0.002). However, this association did not repeat in the following assessments. Most cats (62.8%) kept the same interaction score over assessments. However, among the cats that did change their scores, 77% (n = 10/13) changed to a positive result. There was no difference between males and females in SIC (P = 0.27).

The mean RRs in each of the assessments are shown in Table 3. The mean RR in the cat-specific music group had minimal changes during the assessments. In the classical music group, apparently the mean RR gradually reduced; and in the control group, the mean RR increased on A4. However, only in the A4 group was there a statistical difference in the RR mean between the classical music group and the control group (P = 0.02).

Table 3.

Respiratory rate of cats during the five assessments (A1–A5), according to music group

Assessments	Cat-specific (n= 11)	Classical (n = 12)	Control (n = 12)
A1	28 ± 10.5	34.6 ± 18.2	31.6 ± 9.8
A2	28 ± 7.4	29.3 ± 10	31.4 ± 14.1
A3	28.8 ± 6.7	27.2 ± 5.8	28.4 ± 10.6
A4	28.8 ± 4.8	24.7 ± 4.3*	32.7 ± 9.4*
A5	28 ± 7.6	24.4 ± 3.7	31 ± 9

Data are mean ± SD

^*Significant difference between groups (P = 0.02)

There was no difference regarding voluntary food consumption between groups. Most cats (62.2%) showed appropriate feeding behavior only about 20 h after being hospitalized.

During the analysis of the salivary cortisol, it was noted that only six cats had enough volume of saliva for the two analyses. Enough saliva volume for an individual analysis was collected in 16/35 cats on the first day and in 10/35 cats on the second day. Several samples had blood and food contamination. Most cats rejected swab permanence in the oral cavity to achieve an adequate collection. Some cats bit the swab and, to prevent the risk of ingestion, the collection had to be discontinued in these cats.

Discussion

Many stressful noises can be present in the hospital environment,^23,24 and music may help some animals to cope with stress. ¹⁰ However, to avoid adverse effects, the music volume should be below a conversational level and only during the animals’ active periods. ⁹ In our study, the music was played only during periods when the hospital environment had increased activity (ie, the noisiest times).

Acute stress responses in cats subjected to short-term cage environments are not well known, compared with chronic stress responses in long-term cage environments. ³¹ Cats’ behavioral and physiological coping responses must be better understood to ensure the well-being of cats kept in cages. ³² Stress evaluations involving interaction with cats may be more useful than the CSSs, as the score can underestimate stress, as has already been reported in sheltered cats. ³³ As the first reaction of a cat under stress is to stop interactions, ³⁴ using a tool to recognize this reaction beforehand may be useful, mainly because some shy cats may be more relaxed after a positive interaction with a person. For this reason, we tested a SIC score using described behaviors, in a similar way to Zeiler et al. ³¹ We found a significant difference among groups at the first assessments. The cat-specific music group had a higher percentage of positive interactions with the researcher than the other groups. Our results agree with the observations of Hampton et al, ²⁶ who concluded that handling cats when listening to cat-specific music is easier than in cats listening to classical music or silence. ²⁶ However, in our study, the difference found in social interactions was not replicated in the following assessments. It is hard to imagine whether SIC can be used as an acute stress measuring tool or if it is only a reflex of cats’ temperaments. ³¹ In our study, most cats kept the same interaction score over the assessments. Among the cats that changed their scores, 77% changed to a positive result. The emotional stability in most cats during SIC in our study might be a sign of wellbeing, as there was no negative anticipation, even when cats were handled many times for normal procedures during their period of hospitalization.

There was no difference between sex and age in the CSSs, which differed from a study that showed higher scores in male and older cats. ³⁵ However, Hampton et al also did not find a difference between sex and age in the CSSs. ²⁶ The mean CSS of the five assessments in the groups was between 2 and 3 (ie, between ‘weakly relaxed’ and ‘weakly tense’). These results were similar to findings in another study. ²⁸ Although widely used, there is a lack of studies about the application of the CSS in unhealthy cats, as all studies we analyzed described the use of the score on healthy owned or healthy sheltered cats.^{26,28,33,36–38} It is unknown how pain or disease can influence the CSS measuring system, or how demeanor affects feline pain-assessment scales. ³⁹

Although widely used as a stress-assessment method, there is discussion about the effectiveness of the CSS, as in reality it measures fear, which is different from stress. ³⁹ Furthermore, when using scales based on a cat’s behavior, we must remember that cats’ instincts include hiding signs of illness, pain or stress, ⁴⁰ so the scale may have limitations. In addition, the CSS should not be applied at temperatures <15°C because cats cannot assume a relaxed position at this environmental temperature. ²⁸ In our study, the average temperature in the cat cages was 24.1°C, with a minimum of 19.1°C.

The mean RR in the classical music group on the fourth assessment showed a significant difference compared with the control group (Table 2). Although not statistically significant, the mean RR apparently decreased in the classical music group over the course of all assessments, compared with the control group, while the mean RR in the cat-specific music group remained the same. A study with anesthetized cats showed that animals submitted to classical music demonstrated a lower RR vs silence or other musical genres. ³⁰

Owing to the small number of viable saliva samples and difficulties during sampling, a statistical analysis with salivary cortisol was not possible. Salivation is regulated by the autonomic nervous system and it decreases in response to activation of the sympathetic nervous system; thus, hospitalized cats are expected to produce less saliva when stressed, and this may impair sample collection. ⁴¹ Another study reported difficulties in saliva collection and sample contamination with food and/or blood. ⁴¹ Another limitation in salivary cortisol measurement is the difficulty in achieving a positive correlation between plasma cortisol and cat behavior.^41,42 A study showed a relationship between salivary cortisol with behavioral reactions in dogs, associating stress-related behaviors with increased salivary cortisol. ⁴³ Another study correlated salivary and urinary cortisol with plasma cortisol in dogs, showing a connection between them. ⁴⁴ A study failed to show a correlation between CSS for behavior measuring and physiological measuring in cats, such as cortisol levels. ³³

Some of the limitations of this study included lack of an adequate stress score tool for hospitalized cats. There is currently not an effective non-invasive accurate method to assess stress in cats. ³⁹ Future research may create a wellbeing score for hospitalized cats that includes a pain scale. Another limitation was the difficulty in comparing cats with different diseases, which may affect the way cats react to stress. Also, the number of times cats were being manipulated for medical assessments, treatments and sample collections could affect their stress levels.

The music had good acceptance by most veterinarians and technicians, although some sounds present in the cat-specific music were not pleasing to some people. Music therapy may have a relaxing effect on technicians, which can lead to lower levels of stress in animals during their care. ⁹ This phenomenon may have been a confounding factor in our study.

Conclusions

There was no difference in the mean CSS between groups. Classical music vs control had a lower mean of RR in one of the assessments. Also, cat-specific music presented a higher number of positive interactions between cats and researchers, but only in one of the assessments. As for salivary cortisol, not enough samples were obtained for statistical analysis, suggesting that this method may not be adequate to measure stress in hospitalized cats. Both cat-specific music and classical music seem to be of some benefit to hospitalized cats.