Factors Affecting Sleep Among Dogs and Cats in a Veterinary Intensive Care Unit

ABSTRACT

Objective

To evaluate the amount of sleep obtained by hospitalized dogs and cats in an intensive care setting and to identify factors that may impact veterinary patients’ sleep.

Design

A prospective, observational study spanning a 4‐week period in June of 2020.

Setting

Academic teaching hospital.

Animals

A total of 96 dogs and 16 cats hospitalized in the ICU during the 4‐week study period.

Interventions

None.

Measurements and Main Results

Patient activity was categorized as active, resting, or asleep and was recorded along with ICU environmental data on an hourly basis. Environmental data consisted of subjective assessment of noise level, ambient lighting, number of people present, and number of hospitalized patients. The median observed time asleep was 40% and 11% for dogs and cats, respectively. During natural nighttime hours (9:00 p.m. to 6:00 a.m.), the odds of a patient being asleep were 1.7 times higher if lights were dimmed (p < 0.001). Patients were also less likely to be asleep with higher noise levels (odds ratio 0.66 for each increase in noise level, p = 0.003).

Conclusions

Hospitalized dogs and cats experience sleep disturbances similar to those reported in human ICU patients. Ambient noise and light are significant factors contributing to sleep disruption in cats and dogs hospitalized in a veterinary ICU. The findings of this study support implementing efforts to promote patient sleep through environmental modifications in the veterinary intensive care setting. Additional research is needed to establish objective means of assessing sleep in hospitalized dogs and cats, to determine sleep patterns of hospitalized veterinary patients, and to quantify the impact of sleep disturbances on veterinary patient convalescence.

Abbreviations

OR

odds ratio

1. Introduction

Sleep plays a vital role in normal bodily function and healing, with alterations in sleep potentially affecting patient convalescence in the hospital setting. The impact of disturbed sleep on people in the ICU has been explored extensively over the past four decades [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13]. Consequences of sleep deprivation in human ICU patients include functional disturbances in ventilatory, cardiovascular, immunologic, hormonal, and metabolic homeostasis, as well as an increased risk of psychogenic stress and delirium, all of which influence morbidity and mortality in critically ill patients [6, 7, 11, 14]. Factors contributing to sleep disturbances in human ICU patients are categorized as either intrinsic or extrinsic. Intrinsic factors are patient specific and include severity of underlying disease, medications, pain, anxiety, inflammatory mediators, and circadian rhythm disturbances [4]. Extrinsic factors affecting sleep in hospitalized patients are related to the hospital environment, including noise, light, and nursing activities [4, 9].

Although the importance of sleep and the prevalence of sleep disruption in hospitalized patients are well established in people, there is a limited understanding of normal sleep patterns and the effects of hospitalization on patient sleep among veterinary patients. A review from 1984 estimated that domestic dogs sleep mainly at night and obtain an average of 8.4–12.9 h of sleep daily, with mean sleeping episodes of 45 min, while domestic cats on average sleep 13.2 h daily, with a more flexible sleep cycle and individual sleep periods ranging from 50 to 113 min [15]. Cats demonstrate “erratic nocturnal” sleep patterns more frequently than dogs, with a recent paper documenting owner observations that their cats predominantly sleep more during the day and are more likely to be awake at night [16]. While some studies have characterized baseline sleep patterns in healthy dogs [17, 18, 19, 20, 21, 22, 23], there are fewer documented studies in healthy domestic cats [24, 25, 26]. The authors are not aware of any studies describing sleep patterns in hospitalized dogs and cats. Given the alterations to the nighttime environment in the hospital setting, particularly ambient noise, lighting, and disruption from hospital personnel, it is plausible that the quantity and quality of sleep among hospitalized veterinary patients differ substantially from their normal sleep patterns. For example, two veterinary ICUs were found to have sound pressure levels exceeding World Health Organization guidelines for hospital noise, suggesting that sound in veterinary ICUs is loud enough to disrupt sleep in critically ill patients [27]. The impact of these suspected sleep disruptions, however, has yet to be characterized in the veterinary ICU setting.

The objectives of this observational study were therefore to (1) establish a preliminary understanding of the amount of sleep obtained by dogs and cats hospitalized in an ICU and (2) identify factors impacting their sleep. We hypothesized that veterinary patients in the ICU, similar to human patients [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13], would sleep less while hospitalized than what is expected for nonhospitalized patients. We further hypothesized that noise level and ambient lighting would be inversely related to patient sleep.

2. Materials and Methods

2.1. Data Collection

A prospective, observational study was conducted over a 4‐week period in a veterinary teaching hospital ICU. The recorded total area of the ICU was 1719 square feet. The study was approved by the Institutional Animal Care and Use Committee and included canine and feline patients hospitalized in the ICU for a minimum of 24 h. Data sheets were designed to collect information regarding both patient and environmental variables. Additional variables were collected from the daily patient census and patient treatment sheets. Patient variables included signalment, presenting complaint, number of days in ICU, administration of a sedating drug (including sedative, anxiolytic, or analgesic) or general anesthesia within a 24‐h period, and patient activity level collected hourly. Environmental variables collected hourly included time, kennel location, noise level, ambient light level, and number of people in the ICU. For cats, hourly observations included the use of pheromone spray^a, the presence of hiding boxes in the kennel, and whether the patient was hiding. A comment box was also provided for ICU staff to provide additional information. The total number of patients in the ICU at 11:00 a.m. and 11:00 p.m. was recorded daily.

An ethogram was adapted for evaluating patient activity levels based on a previously described scale [22]. Activity level was classified as active, resting, or asleep (Table 1). A 10‐point subjective noise scale was modified to a 3‐point scale to better correspond to the ICU environment (Table 2) [27]. Light was classified as “bright” or “dim” to reflect whether white, fluorescent lights, or dimmed green lights were the predominant light source. The number of people in the ICU at a given time was recorded as <5, 5–10, or >10.

TABLE 1.

Classification scheme used by clinicians and technicians to characterize activity level of hospitalized dogs and cats in the veterinary ICU during the 4‐week observation period.

Species	Active	Resting	Asleep
Dog	Any activity other than resting or sleeping. Eyes are open .	Dog is lying with its abdomen touching the bottom of the kennel/run, with legs extended, curled close to the body, or laid to one side. Eyes are open .	Similar to resting, but eyes are closed . Possible twitching of paws, ears, whiskers, tail, and eyes and some vocalizations (muffled barks, whines, howling) may also be observed.
Cat	Any activity other than resting or sleeping. Eyes are open .	Cat is lying with its abdomen touching the bottom of the kennel/run, with legs extended, curled close to the body, or laid to one side. Eyes are open .	Similar to resting, but eyes are closed . Possible twitching of paws, ears, whiskers, tail, and eyes and some vocalizations (muffled meows, whines) may also be observed.

TABLE 2.

Classification scheme used by clinicians and technicians to quantify noise level within the veterinary ICU at different time points throughout the 4‐week study period.

1	2	3
Someone could easily hear you use a whisper/very quiet voice from 3 feet away (e.g., library)	You could easily hold a conversation with someone 3 feet away without raising your voice	Conversation is possible with someone 3 feet away, but requires you to raise your voice (e.g., noisy cafeteria)

ICU technicians and doctors were oriented to the study and provided with instructions for how to complete data collection. The intended frequency of observations was once per hour; ICU staff were instructed to record observations for the study while completing their standard hourly patient monitoring. A minimum time between observations was not specified. An initial pilot study was performed over a 2‐week period, after which data collection sheets were modified based on feedback from ICU staff. The 4‐week study was then initiated, with patient and environmental data collected hourly. ICU staff were instructed to collect patient data by observation before treatment administration to minimize the effect of patient care on observations.

2.2. Statistical Methods

Continuous variables were reported as median with range or interquartile interval, and categorical variables were reported as modes. For the purposes of this study, natural nighttime hours were considered to be between 9:00 p.m. and 6:00 a.m.

To examine the effect on sleep of various factors of a patient's stay, a multiple mixed logistic regression model was fit with a binary response of asleep versus active/resting and predictors consisting of light as dim or bright, a numeric noise factor, an indicator variable for whether the patient was asleep at the previous check, the patient's location as numbered cage or other, the patient species, the numeric day of their stay, and time of day as categorical, allowing for interactions between time of day and day of stay. A random intercept was also included to account for individual tendencies regarding sleep. Hours were treated as categorical to allow for any pattern of wakefulness and to allow for any pattern of change through the days.

To examine the effect of length of stay on longest period asleep, the longest sleeping event in each 24‐h period was identified for each patient, with sleep periods that ended before 8:00 a.m. being attributed to the prior day. If a patient's status was not recorded during a check, it was recorded as “not observed” to avoid inflating or underestimating sleeping times. Sedation and anesthesia were also summarized to a per‐day value by transforming to whether there was any time point that day at which they were considered sedated or anesthetized; the anticipated duration of a sedative's effect or general anesthesia's effect was not specifically recorded or evaluated. A linear mixed model was then fit with kennel location, sedation, species, anesthetization, separate indicators for the first day and the last day of hospitalization, and the intervening day(s) of hospitalization as predictors with the length of longest period of sleep as the response with a random intercept for each patient to account for individual tendencies pertaining to length of sleep.

All analyses regarding length of stay were subset to only the first 6 days of patients’ stays to prevent the two patients with stays longer than 6 days from biasing the results. The first and last day of ICU hospitalization were excluded from summary statistics to account for incomplete observation associated with time of ICU admission and discharge. Estimates are reported as odds ratios (ORs), with the multiplicative change in odds based on a change in the variable. The threshold for statistical significance was set at 5%, with p‐values less than 0.05 considered significant.

3. Results

3.1. Descriptive Data

A total of 112 patients (96 dogs and 16 cats) were included in this study. The median age of patients observed was 8 years (range: 4 days to 16 years), and the median weight was 12.3 kg (range: 0.3–47.2). The median number of patients in the ICU at 11:00 a.m. and 11:00 p.m. was 9 (range: 5–14) and 8 (range: 5–15), respectively. Reasons for hospitalization are listed in Table 3. Ninety patients (80%) received at least one sedating drug during their stay, and 18 (16%) underwent general anesthesia.

TABLE 3.

Reported reason for hospitalization for observed canine and feline ICU patients (n = 112) throughout the 4‐week study period.

Type of disease	Specific condition
Cardiovascular disease (20)
	Congestive heart failure (7)
	Dilated cardiomyopathy (2)
	Heart‐based tumor (2)
	Hypertrophic cardiomyopathy (2)
	Aortic thrombus, suspected
	Heartworm infection with caval syndrome
	Hypertension
	Left atrial rupture
	Pericardial effusion
	Pulmonary hypertension
	Third‐degree atrioventricular block with pacemaker placement
Gastrointestinal/alimentary tract disease (19)
	Pancreatitis (7)
	Gastrointestinal foreign body (3)
	Vomiting (2)
	Diarrhea
	Gastric dilatation volvulus
	Gastrointestinal ulceration
	Gastrointestinal lymphoma
	Gastrointestinal obstruction
	Megaesophagus
	Salivary mucocele
Unspecified (19)
	Presenting complaint not recorded
Hematological disease (10)
	Immune‐mediated hemolytic anemia (5)
	Immune‐mediated thrombocytopenia (2)
	Anemia
	Anemia and thrombocytopenia
	Hematopoietic neoplasia
Urinary disease (9)
	Azotemia (2)
	Ureteral obstruction (2)
	Protein‐losing nephropathy, suspected
	Pyelonephritis
	Renal disease
	Renal failure
	Renal injury
Other (9)
	Snake bite (2)
	Alopecia
	Chocolate intoxication
	Hemoabdomen
	Maxillectomy
	Neoplasia
	Septicemia
	Systemic inflammation
Respiratory disease (7)
	Aspiration pneumonia (3)
	Laryngeal swelling
	Lower airway disease
	Myasthenia gravis with aspiration pneumonia
	Pulmonary disease
Hepatic disease (6)
	Hepatopathy (2)
	Liver lobectomy
	Liver mass
	Ruptured hepatic abscess
	Liver failure
Endocrine disease (5)
	Diabetes mellitus (2)
	Insulin overdose
	Primary hyperparathyroidism
	Hypoadrenocorticism, suspected
Musculoskeletal and neurologic disease (4)
	Intervertebral disc disease
	Lameness
	Myelopathy
	Nonambulatory paraparesis
Splenic disease (4)
	Splenic mass (2)
	Splenectomy
	Splenomegaly

Note: Number in parentheses indicates multiple patients with the condition.

Excluding the first and last day of hospitalization, dogs were observed for a median of 9 h a day (quartiles: 5–14 h) and cats for 7 h a day (quartiles: 4–12 h). Patients were found to be asleep a median of 37.5% of the time that they were observed. Dogs were asleep 40% of the time observed (quartiles: 18% and 54%), and cats were observed asleep a median of 11% of the time observed (quartiles: 0% and 19%). For cats, no significant change in observation completion consistency was found at different hours (p = 0.332). For dogs, observations were significantly more consistently completed in the late morning and late evening hours (p < 0.001).

Overall, lights were observed to be bright during 82.5% of observations. Of the 280 observed nighttime hours, 106 observations (37.9%) indicated dim lights and 174 (62.1%) indicated bright lights. Almost all observations with dim light (97.8%) were between the hours of 10:00 p.m. and 6:00 a.m., and dim lighting was never observed between 8:00 a.m. and 7:00 p.m.

Of the noise level observations, 62.5% were classified as level 1, 35.8% as level 2, and 1.6% as level 3. The ICU tended to be louder at 7:00 a.m. (mean 0.31 higher score, p = 0.013) and 8:00 a.m. (mean 0.41 higher score, p < 0.001) or when more people were observed in the ICU (mean 0.14 higher score for 5–10 people and 0.61 higher score for >10 people, p < 0.001 for both categories). The mean noise level at a given hour was also significantly associated with the number of patients in the ICU during that 12‐h period (p < 0.001).

The most commonly observed number of people in ICU was 5–10, with 59.0% of observations reporting 5–10 people. Fewer than five people in the ICU were recorded during 35.4% of observations, and >10 people were reported for the remaining 5.6% of observations.

Pheromone spray and hiding boxes for feline patients were not recorded as being used at any point during the study; thus, these variables were excluded from statistical analysis. The interaction between hour and day of stay was not significant (p = 0.831), indicating that no consistent relationship was found between likelihood of being observed asleep and day of hospitalization; this interaction was therefore removed from the model. Similarly, the number of patients in the ICU did not have a significant effect on activity status, and this was also removed from the model (p = 0.903).

Patients that were observed to be asleep during the preceding observation were 3.1 times more likely to be asleep at the next observation (p < 0.001). Cats were significantly less likely than dogs to be asleep during a given check (OR: 0.44, p < 0.001). Patient sleep and noise level were inversely related, with patients significantly less likely to be asleep with each increase in noise level (OR: 0.67, p < 0.001). Similarly, patients were significantly more likely to be asleep when lights were dim compared with when lights were bright (OR: 1.66, p < 0.001).

Kennel location, administration of a sedating drug, and administration of general anesthesia were not found to significantly impact the quantity of sleep in a day or the longest period of sleep in a day (p = 0.904, 0.326, and 0.526 and p = 0.988, 0.217, and 0.342, respectively). Cats’ longest period of continuous sleep was an average of 1.1 h shorter than that of dogs (p < 0.001).

Controlling for number of hours observed, patients were found to have more sleep overall with increasing day of hospitalization, with patients sleeping an average of 0.4 h more each day that they were in the ICU (p < 0.001). Patients tended to sleep 0.7 h less on the first day of ICU hospitalization compared with other days (p = 0.026), with their period of longest sleep increasing by 0.2 h per day with increasing day of hospitalization (p = 0.003). However, patients’ cumulative sleep on their final day of hospitalization and period of longest sleep did not significantly differ from preceding days (p = 0.810 and p = 0.860, respectively).

4. Discussion

This study is the first to evaluate the amount of sleep among hospitalized dogs and cats and to describe the factors contributing to sleep disruption in the veterinary ICU. The collected data suggest that sleep in dogs and cats hospitalized in an ICU setting is markedly altered, similar to what is reported in critically ill human patients [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14]; environmental factors including noise level and brightness in the ICU were inversely associated with patient sleep.

Importantly, this study was performed in a single veterinary teaching hospital ICU with a unique environment; the degree to which the findings of this study can be extrapolated to other veterinary ICUs is unclear and requires further study. Specifically, the ICU evaluated in this study is a single‐room ICU housing up to 24 cats and dogs with no separate housing for feline patients. The ICU is distinct from the emergency department treatment room, such that treatments performed in the ICU are limited to those performed on ICU‐hospitalized patients (emergency triage patients are evaluated and treated in a separate room). Major treatment times are scheduled at 7:00 a.m., 1:00 p.m., 7:00 p.m., and 1:00 a.m.; as a result, these and the preceding hour are generally the busiest in the ICU, and there are no specifically designated quiet hours.

At the time of this study, there was no policy or guidelines on when ICU lights should be dim. Upon discussion with the ICU staff, some technicians expressed a preference for dimmed lights during overnight hours but also indicated that certain staff preferred bright lights overnight; thus, the overnight light status was generally a reflection of which technicians were working the overnight shift. It is unknown whether technicians may have dimmed the lights in response to observing patients sleeping.

The percentage of time spent asleep by the dogs and cats in this study suggests that hospitalized small animal patients in the intensive care setting may sleep less than the average 10.1 h reported in healthy, nonhospitalized dogs and the 13.2 h reported in normal cats [15, 20]. Potential explanations for the observed sleep differences in veterinary ICU patients may be extrapolated from documented factors impacting sleep in critically ill human patients [3, 9]; this is supported by the findings that noise level and ambient light were associated with altered patient sleep in the current study. Due to the limitations imposed by inconsistent observation of ICU patients during this study, the authors recommend appointing a designated observer for future observational studies to improve the consistency of data collection and to minimize bias.

Hospitalized dogs and cats were consistently surrounded by other patients in the ICU, with a median of eight to nine patients recorded at representative sample times. Although the exact patient number is expected to vary among institutions, the basic principle of shared space and communal housing is ubiquitous in the veterinary ICU setting. This study found a significant effect of the number of patients in the ICU on sleep status, supporting the argument that higher numbers of patients increase sleep disruption in the small animal ICU, likely due at least in part to an increase in ambient noise. This finding also supports the idea of segregating large ICUs into smaller rooms when feasible to reduce noise exposure, particularly for cats, which are likely to be more negatively impacted by dogs vocalizing [29].

Noise level and room brightness were found to significantly impact sleep in dogs and cats hospitalized in the ICU. Consistent with previous findings [27], our ICU was a noisy environment, with ambient noise climbing to a level 2 out of 3 approximately one third of the time. Unsurprisingly, the loudest periods were observed when more people were present in the ICU or during major treatment hours when students, house officers, and nursing staff were actively working with patients (most notably 7:00 a.m. and 8:00 a.m.). Lights were observed to be bright for more than 80% of the time in a typical 24‐h period. Patient sleep was inversely associated with environmental noise and brightness, with patients less likely to be asleep when noise level was elevated or lights were bright. Whether animals are naturally more awake at times with increased noise level and bright lights, or whether they are awake due to the disruption caused by treatments and activity in the ICU requires further study. These findings are consistent with prior reports of sleep disruption in human ICU patients [2, 3, 6], suggesting that treatment‐related disruption may also play an important role in the veterinary ICU. One feasible intervention to promote patient sleep and potentially improve convalescence is to implement designated “quiet hours” and automated light settings to minimize ambient noise and light between the hours of 9:00 p.m. and 6:00 a.m. to correspond with the normal canine sleep cycle [20]. Lefman et al. describe additional strategies to support sleep in hospitalized veterinary patients, including reducing ambient noise and light, minimizing disruptions associated with nursing interventions, and proper use of sedation, all of which align with the findings of the current study [30]. Also, pheromone spray and hiding boxes for feline patients were not routinely used in our study; optimizing the use of these relatively accessible and inexpensive tools may provide additional methods for reducing stress and improving cats’ sleep during ICU stays.

Although both species were susceptible to sleep disruption in the ICU, feline patients appeared to be more significantly affected compared with canine patients. Hospitalized cats were less likely than dogs to be sleeping at each time check. The classification scheme used in this study also fails to account for the possibility of feigned sleep, which is a described behavior in stressed cats and inherently difficult to distinguish from true sleep [31]. Therefore, estimations of feline sleep in this observational study could have been inflated by feigned sleep, suggesting that hospitalized cats may sleep even less than we have reported. The observed species difference in likelihood of being observed asleep is particularly striking given that cats are reported to sleep more than dogs under normal conditions [15]. The increased sleep disturbance among cats in this study is unsurprising given the previously described differences in susceptibility to psychogenic stress in hospitalized cats vs dogs [30, 32]. This finding emphasizes the particular need for specific interventions to reduce stress and promote sleep among hospitalized cats. Creation of a designated area to allow for separate ICU housing for cats and dogs could reduce ambient noise levels and stress for cats and thereby help to promote sleep. Alternatively, Stoneburner et al suggest that stressed cats may benefit from having a tinted plexiglass sheet at the front of their cage to curtail visual and auditory stimuli without completely eliminating the staff's ability to visualize the patient [32].

The period of longest sleep for veterinary patients in this study increased with increasing day of hospitalization. This finding suggests that hospitalized dogs and cats may be capable of acclimating to the hospital environment over time. However, acclimation required at least 24 h, which is arguably enough time to accumulate significant sleep debt that could impact patient health. Further study of the impact of different techniques for minimizing patient acclimation time to the ICU is indicated. Interventions may include environmental alterations such as housing cats and dogs separately, providing comfortable enclosures with ample bedding and hideaways, instituting designated “quiet times,” organizing enrichment by means of music therapy or animal‐oriented television, and implementing volunteer programs through which physical comfort and affection are provided to amenable patients for longer durations of time without overburdening the nursing staff [29, 30].

This study provides preliminary data regarding the amount of sleep obtained by dogs and cats in the ICU and evaluates factors that potentially limit patient sleep in this setting. Limitations of this study, including the subjectivity and consistency of observations and the accuracy of using hourly observations to represent continuous behavioral patterns, highlight the need for additional research to better understand and improve the sleep behavior of hospitalized cats and dogs. An important limitation of this study was the use of subjective measures to assess key variables including sound, brightness, and sleep activity. These subjective measures were chosen over more objective measures of sound, brightness, and patient sleep status to facilitate ease of observations and maximize compliance with ICU staff making the observations. Use of subjective measures and ICU nursing staff for observations may have contributed to potentially high degrees of interobserver variability and inaccuracy in estimating actual values. Use of designated observers and objective measures of sound and brightness is recommended for future studies. Additionally, although ICU staff were encouraged to record their observations for a given patient with approximately an hour of intervening time, the absence of a specified minimum and maximum time between hourly observations means that “hourly” observations for a patient could have been made as close as 1 minute apart, or nearly 2 h apart.

Using objective measures to evaluate patient sleep status may help distinguish between asleep and sedate patients and help further assess sleep quality among hospitalized dogs and cats in the ICU. We did not find an effect of sedative drug or general anesthesia administration on patient sleep in this study, although this may represent a type II error. The prevalence and diversity of drugs with a sedative effect administered to patients in this study presented a challenge in assessing how anesthesia or sedation administration affects the likelihood of a patient being observed asleep, particularly with regard to the duration of effect of a particular drug. A broader study encompassing larger subgroups receiving different types of sedation would be better powered to evaluate such effects. Evaluating the discrepancy between true patient sleep and staff perceptions of patient sleep in the veterinary environment, although of great interest, proves challenging due to the difficulty in objectively measuring sleep quality and quantity. This challenge has been described by sleep researchers in the human intensive care setting. For example, Aurell et al. concluded that estimated sleep times in hospitalized people were regularly overestimated by nursing staff when compared with corresponding polygraphic sleep recordings [1]. Logistics of shift changes and the need for 24‐h monitoring also required that data collection be performed by several different ICU personnel with varying levels of experience and enthusiasm for implementing the study protocols.

Because this study was implemented during a period of limited staffing during the COVID‐19 pandemic, understaffing in the ICU also contributed to decreased compliance in recording sequential hourly observations as ICU personnel prioritized performing patient treatments over recording observations if time per patient was limited. The limited number of observations made per patient per day prevented extrapolation to estimate the total amount of sleep obtained per 24‐h period. Additionally, for the sake of tracking continuous sleep, patients were not assumed to be asleep during periods without observation, thus leading to a more conservative estimate of continuous sleep time. This may bias the findings of this study toward underestimating the true amount of sleep obtained by the patients hospitalized in this ICU. As mentioned above, such bias could be mitigated by the use of designated observers and objective assessments. ICU caseload during this time was also lower than pre‐COVID‐19 levels; thus, the findings of this study may not be representative of busy veterinary ICUs during peak times. Because the first and last day of ICU hospitalization were excluded from analysis, the results reported here may also underrepresent the sleep of relatively healthy ICU patients that are only hospitalized briefly in the ICU before being discharged or transferred to a “step‐down” ward, as well as more severely ill patients that do not survive to be hospitalized for multiple days.

As the current study provides an initial evaluation of patient sleep and sleep disruption in one ICU, broader information regarding sleep among dogs and cats hospitalized in this and other veterinary ICUs remains to be elucidated. The disruptive factors identified also warrant evaluation in other hospital wards. The less intensive nature of various “step‐down” wards may facilitate quieter environments with fewer interruptions for treatments and patient evaluation. Thus, another strategy for improving sleep among hospitalized dogs and cats could include de‐escalating their hospitalization to less intensive care wards as soon as medically appropriate, while still prioritizing species segregation where possible to minimize the disproportionate sleep disturbances in cats.

In conclusion, hospitalized dogs and cats appear to experience sleep disturbances in the ICU setting. Ambient noise and light are significant factors contributing to sleep disruption in these patients. Further study is indicated to better define patient sleep patterns among dogs and cats in veterinary ICUs. Additional research is also needed to establish objective means of assessing sleep in hospitalized dogs and cats and to quantify the impact of sleep disturbances on their convalescence. Such measures may also prove useful in evaluating strategies to mitigate hospitalization‐induced sleep disruption in these patients. Extrapolating from data documenting the negative impacts of ICU sleep disruption on hospitalized people, the findings of this study support implementing efforts to promote patient sleep through environmental modifications in the veterinary intensive care setting.

Author Contributions

Study design was developed by A.V.E., V.F.S., A.M.L., M.E.G., S.J.L., and J.B.R. Data collection and synthesis were completed by A.V.E., S.J.L., V.F.S., and E.A.D., and data analysis was performed by A.V.E., J.B.R., and V.F.S. Manuscript composition was performed by E.A.D., A.V.E., and V.F.S. Manuscript editing and review were performed by A.M.L., M.E.G., S.J.L., J.B.R., E.A.D., and V.F.S. The final manuscript was approved by all authors.

Conflicts of Interest

The authors declare no conflicts of interest.