The impact of light,noise,cage cleaning and in-house transport on welfare and stress of laboratory rats

Light and the rat's eye

Although several factors contribute in determining the extent of light-induced retinal damage (light intensity and duration of light exposure; wavelength of light; rat age, strain, and ocular pigmentation; light exposure history of the animal), retinal damage has been reported in albino rats at light intensities greater than 60 lx. ^23,25,26 A preference test system has shown that, even though the effect was more pronounced in the albino rats, both albino and pigmented rats preferred cages with a low light intensity (< 100 lx) over those with higher light intensities (100–380 lx). ²⁷ Using avoidance tests, Schlingmann et al. ²⁸ have shown that albino rats avoid light intensities as low as 25 lx and pigmented rats from as low as 60 lx, but animal rooms need to have enough light for the technicians to perform animal care tasks. A minimum of 210 lx at working height as reported by the same authors has sufficient light in the room for the health and good performance of personnel. ²⁹ Providing rats with a place to hide from too much light is therefore strongly advisable.

The cyclic (12 h light, 12 h dark; 12L:12D) light intensities under which the rats are reared have been shown to influence the rod outer segment length, photoreceptor cell density, whole retina visual pigment regeneration rates, the concentration of fatty acids and cholesterol in photoreceptor outer segment membranes and antioxidant capabilities of the rat retina. The retinas of rats reared under high-intensity cyclic light (i.e. 400 lx) acquired characteristics that made them less susceptible to light damage when compared with rats reared under low-intensity cyclic light (i.e. 6 lx). Such characteristics included reduced numbers of photoreceptor cells, reduced outer segment length and pigment concentration, decreased levels of polyunsaturated fatty acids and increased antioxidant capabilities. ^23,30,31 Light-induced retinal damage is triggered by rhodopsin bleaching and albino rats have been shown to adapt to changes in environmental light intensities by adjusting the amount of rhodopsin per retina. ^32,33 Case and Plummer ³⁴ reported that the adult retina responds to a different lighting environment by a relatively rapid change in the size of photoreceptor segments, by a progressive and large change in number of ribbon synapses and by a slower progressive and large change in the size of photoreceptor nerve terminals.

Light history experienced by the animal also influences retinal response to light. Dim-light-reared rats have been shown to exhibit an age-related increase in retinal light damage susceptibility, whereas dark-reared rats were equally susceptible to damage at all ages. ^26,33

Retinal damage has been shown to be greater in rats exposed to a dose of intense light during their night period rather than in rats exposed to the same light treatment during their day period (circadian-dependent retinal light damage). ³⁵ The mechanism by which susceptibility to retinal light damage in rats occurs seems to involve a light-induced apoptotic process of visual cell death. ^26,33,35 A single, relatively short, intense light exposure can cause a circadian-dependent, oxidatively induced loss of photoreceptor cells. ³⁵ Ideally, laboratory rats should be kept in the same light cycle and intensity for their life period in the laboratory in order to avoid physiological changes and damage as those mentioned above. Maintaining a standardized light environment of the animals would contribute to their welfare and to the reproducibility of experimental results.

Light, behaviour and physiology

Animal activity and behaviour are also influenced by environmental light intensity. Defecation rate in hooded rats (PVG/C) can be increased under bright lighting (given by a white 150 W bulb suspended centrally on top of a Y-maze placed in a grey-curtained enclosure of 1.37 × 1.37 × 1.22 m high). ^10,36 Intense light conditions (572 lx) markedly suppressed social play behaviour of Wistar Han male juvenile rats accustomed to dim light conditions. In the intense light test conditions, pinning and boxing/wrestling was absent and following/chasing behaviour was markedly decreased. ³⁷ These characteristic social play behaviours are important for the animals' (social) development, and the animals will learn a task to obtain the opportunity to play. Rats preferably play in sheltered places, ³⁸ which are nearly absent under intense light conditions. ³⁷ The presence of enrichment objects providing shelters can therefore be important, creating areas with reduced light intensity inside the cage and enabling rats to hide.

It is difficult to separate the effects of light intensity from those caused by different light wavelengths and there are not many scientific data available on the effect of the colour of light on laboratory animals and even this is occasionally conflicting. ^10,39 However, red fluorescent light has been shown to still serve as a synchronization pulse for rats kept in the dark, ⁴⁰ even with light intensities below 1 lx. ^39,40 Time of ovulation has been shown to be shifted by red light, which suggests that the circadian rhythm of luteinizing hormone (LH) secretion could be entrained by red light. ⁴⁰ However, red light is still being used as dark (e.g. in behavioural studies). In mice, voluntary wheel-running activity has been shown to be influenced by differently coloured lights. ⁴¹ Red light might have an effect on the animals' behaviour and physiology, which made Wersinger and Martin ⁴² advise researchers performing studies on social behaviour not to assume that laboratory rodents are unable to detect dim red light and to minimize the exposure of subject animals to all sorts of light during the dark period.

During light periods, the fluorescent light that is commonly used in animal facilities provides a narrower range of wavelengths than sunlight; therefore, using full spectrum fluorescent bulbs could be considered a way of providing a more natural light environment in the laboratory.

Light is the most important environmental signal regulating the temporal pattern of animal behaviour and physiology, regulating circadian rhythms and stimulating and synchronizing breeding cycles. ¹⁰ The hypothalamic suprachiasmatic nucleus (SCN) contains the primary circadian oscillator in mammals, and SCN neuronal activity is directly regulated by environmental light via the retinohypothalamic tract (pineal/pituitary/hypothalamic-neuroendocrine pathways, starting from retinal photoreceptors). ^10,43,44

Breeding cycles

The duration of the oestrus cycle in rats has been shown to be affected by changes in the proportions of the light:dark (L:D) regime used. ¹⁰ Time of ovulation and gestation length in rats are also influenced by the L:D cycle under which animals are reared. ^10,45 Nocturnal exposure of female Sprague-Dawley rats to light of minimal intensity has been shown to produce a substantial incidence of ovarian changes, suggesting that the incidence of ovarian atrophy observed by the authors in a previous study may have been due to transient exposure to indirect nocturnal light of minimal intensity. ⁴⁶

Circadian rhythms

Continuous exposure to bright light has been shown to strongly suppress circadian rhythms of the sleep–wake cycle, drinking, locomotion, blood pressure, heart rate (HR) and body temperature (BT), ^47–49 which are regulated by the SCN. ^43,50 Constant light (LL) housing has been shown to increase corticosterone levels in both male and female rats, and is used as an experimental model of chronic stress. ^51,52

As mentioned above, in adult rats, after a long exposure to LL, the circadian periodicity disappears, as do rhythms of motor activity, BT, ^47,53 plasma melatonin ⁵⁴ and sexual hormones. ⁵⁵ However, if previously subjected to LL during their lactation period, adult albino rats have been shown to exhibit a circadian rhythm of motor activity. ^56,57 A critical period for sensitivity and ability to adapt to external factors seems to be located in the middle of the lactation stage of rats. ⁵⁶ Light received during lactation affects the strength of the circadian pacemaker and its sensitivity to light. ⁵⁸ Lighting conditions to which newborn animals are exposed are of great significance, as they will affect the circadian system and condition the further adaptation of the adult animal to the external conditions in which it lives.

The period length of the light/dark cycle under which rats are kept has also been shown to influence the body weight and food intake of young male Wistar rats. ⁵⁹ A shift in the light cycle (reducing its photoperiod or changing the time at which lights are on and of) induces a shift of the circadian rhythms of blood pressure, HR and spontaneous locomotor activity in freely moving rats, which can take up to a week to fully synchronize with the new light cycle. ^60,61

Individually housed Sprague-Dawley and Spontaneous Hypertensive rats implanted with a radiotelemetry transmitter presented reduced basal HR when housed under 10 lx illumination or an 8:16 h L:D photocycle with 200 lx illumination, when compared with controls housed under a 12:12 h L:D photocycle with 200 lx illumination. Even though the pattern of effects varied between strains and between male and female rats, Azar et al. ⁶² concluded that housing rats under 12:12 h L:D, 200 lx ambient light conditions was potentially stressful. Rats' adaptation to a photoperiod reduction seems better when extending symmetrically the darkness into dawn and dusk. ⁶⁰ The use of dimmers in rat rooms to create twilight periods between the light and dark cycles can be recommended ⁶³ even though not much information is available comparing the physiological and behavioural changes in rats after sudden turn off of lights with gradual transition from light to dark and vice versa. Timers providing a gradual light transition would better mimic natural light at dawn and dusk, which have been suggested to be preferable in studies of social behaviour of crepuscular species. ⁴²

Effectiveness of light stimuli on daily circadian clock resetting can be modified by learning and events in the environment that reliably precede the onset of light. ⁶⁴ Light stimuli can gain negative emotional significance through repeated pairing with aversive events, and consequently fail to provide the optimal stimulus for entrainment of circadian rhythms. ⁶⁵

In summary, environmental light is a fundamental factor with an important impact on the laboratory rat physiology and behaviour and should be a well-controlled factor in housing conditions. The housing photoperiod should be stable contributing to both animal wellbeing (i.e. avoiding distress, behavioural disturbances, retinal damage and circadian rhythm suppression) and good experimental results. Albino rats avoid areas with light levels over 25 lx, and as recommended in the European guidelines (European convention for the protection of vertebrate animals used for experimental and other scientific purposes – Appendix A, Council of Europe Convention ETS 123), the animal room should have a low light intensity (enough for the performance of husbandry procedures and inspection of animals, ² e.g. around 210 lx at working height) and darker areas or hiding places should be provided inside the cages (e.g. tubes or shelters to provide rats with some control over their exposure to light). Cages on top shelves might be exposed to too much light and deserve special attention.

Auditory sensitivity

Rodents have a different spectrum of audible sounds with maximum sensitivity at frequencies that are inaudible to humans. Humans can perceive frequencies from 20 to 20,000 Hz, the frequencies from 400 to 4800 Hz being important for speech. Rodents not only produce sounds that we can hear, but also produce and hear frequencies that are inaudible to humans (above 20 kHz), perceiving sounds up to 80 kHz. ^68,70–72 Different methods have been used for determining the auditory range of laboratory animals and the results presented in the literature are diverse. ^66,71,73 Using a combination of an operant conditioning and the psychophysical method of constant stimuli, Gourevitch and Hack located the frequency region of greatest auditory sensitivity for the rat at approximately 1 octave wide in the vicinity of 40 kHz (frequency at which the rat is most sensitive). In this study, water-deprived Wistar rats were trained to respond to a tone by pressing a bar to obtain water as a reward (operant conditioning). The method of constant stimuli was applied by randomly varying the intensity of the signal from presentation to presentation among five intensities chosen from preliminary measurements so that the strongest signal was almost always audible to the rat and the weakest signal was only rarely audible. ⁷⁴ Kelly and Masterton showed a range of hearing from 250 Hz to 80 kHz at 70 dB (SPL) for the (Sprague-Dawley strain) albino rat by using the conditioned suppression technique: first, the animals were water-deprived in their home cage and then trained to lick a spout for water reinforcement in the test apparatus; after this, animals were further trained to associate the offset of a pure 10 s tone with a foot shock until the tone elicited a freezing response incompatible with licking – this freezing or suppression of licking was then used as an indication of the animal's ability to hear a tone. The behavioural tests have shown that the absolute auditory sensitivity of the rats as less than 10 dB SPL at 38 kHz. ⁷² Using behavioural and electrophysiological techniques, Borg ⁷⁵ showed the highest degree of normal auditory sensitivity of albino rats to be around 12–24 kHz, the upper hearing limit being 50 kHz. The behavioural audiogram of the hooded Norway rat was determined for frequencies from 250 Hz to 70 kHz. At a level of 60 dB SPL, the low-frequency limits are 530 Hz for the hooded rats and 400 Hz for the albinos. At the other end, the high-frequency limits are 68 kHz for the hooded rats with an estimated 76 kHz for the albino rats. No effect of albinism was detected by Heffner et al. ⁷⁶ since there were no differences in the audiogram between albino and pigmented Norway rats (Rattus norvegicus). The audiogram of rats, compared with that of humans, is characterized by a lower capacity for detecting low frequencies (under 500 Hz) and a better capacity for detecting high frequencies (over 8000 Hz). ⁶⁸

There are differences in the structure of different cell types in the inner ear between albino and pigmented R. norvegicus, and differences in auditory sensitivity have been related to age, strain or stock differences. ^{67,68,71,76,77} For example, the rat strain Fisher 344 (F344) has a very different auditory sensitivity when compared with the F1 hybrid cross between the F344 and Brown Norway rat (FBN). F344 rats show approximately 20 dB better hearing at low frequencies (4 kHz), whereas FBN rats show approximately 20 dB better hearing at higher frequencies (32 kHz). ⁶⁸ Age differences in hearing ability of rats have also been reported, the first auditory nerve–brainstem-evoked responses have been detected to begin in 7–8-day-old postnatal rat pups. The rat auditory sensitivity increases with the opening of the meatus of the ear at days 12–14 of age and reaches adult thresholds at about 20–22 days. ^67,78 The inner ear of rat pups has been shown to have a maximum susceptibility to acoustic trauma at 22 days of life, after exposure to a continuous 120 dB SPL white noise for 30 min. ^79,80 Auditory deprivation can also have profound effects on the hearing ability of the animal. Studies have suggested a sensitive period, from postnatal days 10 to 16 (including meatus opening), during which proper development of parts of the auditory system of the rat (namely, the dorsal and ventral cochlear nuclei) may require acoustic stimulation. ⁸¹ Poon and Chen ⁸² showed that exposure to trains of tones improved the ability of the exposed animals to process those same tones.

Vocalization/communication

Of course, rats not only react to but also transmit sounds. Ultrasound emission is an important communicating pathway used by rats in several situations, but they can also vocalize below 20 kHz. Ultrasounds were detected during aggressive interactions, sexual intercourse, mothering and during stressful situations (e.g. handling, pain or presence of a predator), and vocalization below 20 kHz were found in aggressive interactions, during the first days of life or while in pain. ^66,67,69,83

Most of the ultrasounds are emitted at a frequency of 21–32 kHz and are therefore called 22 kHz calls. ^67,84 Sales ⁸⁵ has described two major ultrasonic vocalizations during aggressive interactions between male rats: a short signal ranging from 40 to 70 kHz and a long one ranging from 23 to 30 kHz. Observing rats' behaviour, Sales proposed that long vocalizations were indicating submissive behaviour (by the intruder), while short pulses represented aggression from the resident male. Some studies report the emission of 22 kHz calls associated with a decrease in the aggressive behaviour response of the resident rat and in inhibition of attack, but in other studies the submissive rat was found to emit both long (22 kHz) and short (50 kHz) calls and some groups have been unable to find a correlation between the inhibition of aggressive behaviour and ultrasonic vocalization. ^67,85 Female rats were also shown to emit both high- (32–60 kHz) and low-frequency (20–32 kHz) ultrasonic calls in agonistic encounters, with the rate of high-frequency calls enhanced during oestros. Low-frequency ultrasounds were shown to be shorter in duration and higher in frequency than those emitted by male rats in similar conditions. ⁸⁶

Groups of three male and two female Long-Evans rats, living in a visible burrow system, showed high levels of 18–24 kHz ultrasonic cries when presented to a cat (predator). These cries were emitted both during cat presentation and for 30 minutes following removal of the cat, with the rats preferentially staying in the tunnel/burrow system (‘hiding’). In the same study, when single-housed rats were confronted with the cat, either with or without a place of concealment, almost no vocalizations were detected. Blanchard et al. ⁸³ suggested that the production of ultrasonic vocalizations during and after exposure to a predator is greatly facilitated by the presence of familiar conspecifics, and may serve as alarm cries.

Both male and female rats emit ultrasounds during copulation. The sexual ultrasonic calls were found to be divided into 22 and 50 kHz calls. The 22 kHz calls are emitted pre- and postejaculation. ^87,88 Songs of male rats were described by Barfield and Geyer ⁸⁷ as pulses of 1–3 s long and regular, 22–23 kHz in frequency and an intensity that can increase up to 80 dB. Female rats have been shown to produce short (10–200 ms) 40–70 kHz vocalizations during copulation. ⁶⁷ Female rats emit different ultrasounds during the entire oestrus cycle. In response to a devocalized sexually experienced male, scent marking and 50 kHz ultrasonic mating vocalizations of Long-Evans female rats were shown to change in frequency across the cycle, both behaviours being highest at proestrus/early oestrus, which might indicate vocalization as an important factor coordinating reproduction. ⁸⁹ Male pre-ejaculatory 22 kHz calls have also been suggested to be important for inhibition of female agonist behaviour and for stimulation of the female sexual responsiveness. ⁹⁰

Pregnant females emit more ultrasound than non-pregnant females and a diurnal variation has been shown for both, more sounds being emitted during the dark than during the light period. ⁹¹ Mother–pup relationship involves the emission of sounds, from the audible to the ultrasound range. Pups emit ultrasound distress calls in response to cold, isolation or unusual tactile stimulation. ⁶⁷ Calls are emitted by rat pups when they are manipulated or moved by the mother. Unusual tactile stimulation can also induce vocalization. During handling, pups' calls are audible from day 2 onwards, later consisting of both audible sounds and ultrasounds, and at day 10 of only ultrasounds. ⁹²

When analysing the response of rats to two types of ultrasonic vocalizations (50 and 22 kHz) in an emergence test, Burman et al. ⁹³ suggested that rats use different ultrasonic vocalizations to indicate either positive emotional state (50 kHz) or negative emotional state (22 kHz). Rats that received playback of the 22 kHz ultrasonic vocalizations showed an increased latency to emerge and spent less total time in the open arena than rats receiving playback of background noise, suggesting a state of increased anxiety; therefore, the authors suggested that 22 kHz vocalization could induce a negative emotional state in the rats hearing it and could therefore be useful as a welfare indicator for group-housed rats, including both callers and non-calling group mates. ⁹³

Because of its short wavelength, ultrasound fails to penetrate physical barriers well and do not pass through a standard plastic cage, which might protect rodents from ambient ultrasound (depending on the cage structure), except during cage changes or other procedures involving opening the cage. ⁴²

The emission of sounds by rats under stressful situations has also been reported. For example, ordinary handling of rats elicits both audible and ultrasonic calls. ^67,94 Ultrasonic calls were detected when male rats were gently stroked on the head, neck, side of trunk or tail (with the most reactivity being shown when handling the dorsal neck and the least sensitive area being the tail). ^67,94 The response to human handling consisted of multiple series of long 22 kHz calls. The number of ultrasound vocalizing (with the range from 21 to 32 kHz) Wistar Han rats in response to hand touch was shown to be significantly influenced by their housing condition – single versus community cages. Only 60% of the grouped-housed rats emitted ultrasonic calls during the first session of brief hand touch, compared with 100% of the single-housed rats responding to the same tactile stimuli. Rats became quickly habituated from session to session, until extinction of response. Brudzynski and Ociepa ⁹⁴ suggested this quick habituation indicates that the distress response – ultrasonic vocalization of rats – could be caused by a potential danger or threat to the animal and it does not necessarily reflect physical discomfort or pain. Pain-induced vocalizations by laboratory rats are also reported in the literature. For example, electrical stimulation of tail or foot elicited both audible and ultrasonic vocalizations at 20–35 kHz. ^95,96 Acoustic stimuli, which induce a startle response, were also shown to evoke ultrasonic vocalization in the rat. Startle-inducing acoustic stimuli evoked continuous ultrasonic calling, with 50–70% of male Wistar rats emitting long (0.5–1.2 s) ultrasonic vocalizations (22 kHz call). Comparing the development of vocalization behaviour of those rats with the development of freezing in experiments of fear conditioning, Kaltwasser ⁹⁷ concluded that startle-eliciting stimuli may induce a state of fear in the rat.

Auditory effects of sound

Regarding noise influences on laboratory animals' physiology and behaviour, both auditory (hearing damage) and non-auditory effects have been reported and revised in the literature. ^{66,67,69,70,98,99} Intense noise exposure can damage the cochlea and inner ear and lead to a cascade of auditory effects along the entire central auditory cascade. Susceptibility to noise-induced hearing loss is species/strain dependent, and it has been shown to be genetically determined in inbred strains of mice. ⁶⁸ Exposure to uniform stimulus patterns may lead more readily to hearing loss, whereas exposure to irregular patterns may be more likely to cause disorders due to repeated activation of the neuroendocrine system. ⁹⁸ As previously mentioned, rat pups are most sensitive to auditory damage before 22 days of life, during the period of anatomical differentiation of the ear structures. ⁸⁰ Age-related deterioration of hearing function has also been reported in rats, not only as a natural/physiological consequence of aging but also resulting from noise stimuli. ^100,101 Hearing loss in aged rats has been reported in about 20–25-month-old rats, with the largest hearing loss reported between 24 and 40 kHz. ^67,77,101 Constant white noise might also have negative consequences for the normal development of the auditory system of the animals by effectively masking out the normal input to the ear from vocalizations and other sources. ⁶⁸

Audiogenic seizures are another possible auditory effect of sound. An animal with these sound precipitated convulsions might crouch, shiver and indulge in substitute behaviour (e.g. grooming) after a sound stimulus. This stage is immediately followed by uncontrolled running and convulsion involving tonic and clonic episodes. ¹⁰² Audiogenic seizures can produce stressful effects in rats. D'Amour et al. showed that the induction of multiple audiogenic seizures (15–20 in 1 day) had an effect on rats' physiology producing increases of adrenal weights measured as ratios of body weight. The same was not true with the induction of a single seizure. ¹⁰³ In a study by Duncan, the adrenals of albino rats killed after single or repeated one-a-day audiogenic seizures were weighed and analysed for ascorbic acid, cholesterol and corticosteroids. Adrenal hypertrophy and an increase in corticosteroids were shown as response to repeated stress and decreases in cholesterol and ascorbic acid and increase in corticosteroids were demonstrated as acute response to audiogenic stress. ¹⁰⁴ Rats' ‘hyperbaric oxygen seizures’ have been shown to be caused by the loud hissing of oxygen as it entered a chamber where the animals were placed, in combination with the rough handling of those animals, being in fact a consequence of noise and handling stress. ¹⁰⁵ There are differences on susceptibility to audiogenic seizures among different strains of rodents (genetically-determined susceptibility), but non-susceptible strains can be made audiogenic seizure-susceptible by exposure to particular stimuli during a critical period of postnatal development. Acoustic priming of Long-Evans rats with intense sound during postnatal development subsequent to auditory function was shown to generate audiogenic seizure susceptibility. The rats were exposed to 125 dB SPL 10 kHz tone bursts at 14–36 days of age and tested with white noise at 14 or 19 days following sound exposure. All priming/testing combinations yielded audiogenic seizure susceptibility. All subjects displayed clonus at testing intensities of 120 dB, although some seizure behaviours could be elicited at 100 dB. Repeated testing at 120 dB increased latency to clonus and clonus duration, and total wild running activity. ¹⁰⁶

Non-auditory effects of sound

Several non-auditory effects of noise are described in the literature. The noise level in animal houses and laboratories can be sufficient to act as stressors. Noise activates the sympathetic division of the autonomic nervous system, producing a stress response with physiological characteristics similar to those triggered by other sensory stimuli. ⁶⁸ Intense noise can cause alterations in gastrointestinal, immunological, reproductive, nervous, and cardiovascular systems, blood cell counts, as well as changes in development, hormone levels, adrenal structure, metabolism, organ weights, food intake, body weight and behaviour. These findings have been reported and revised by several authors. ^{10,66–68,70,98,99} Below are some examples of non-auditory effects of noise in laboratory rats.

A decreased body weight gain and food intake was shown in Wistar male rats submitted to noise stress (15 min daily exposure to an acoustic tone of 2640 Hz, 30 W). ¹⁰⁷ Wistar rats exposed, in an acoustic chamber, to 1.5 h of white noise per day at intensities of 107–112 dB displayed increased adrenal weights when compared with the control condition of 60 dB (background noise level). High-intensity noise-exposed rats had increased total leukocyte counts and a relative eosinopaenia. During the six days of noise exposure, noise-exposed rats gained about 15% less weight than controls. ¹⁰⁸ Increased vasoconstriction and increased HR have been reported as non-auditory effects of sound in rats (revised in Turner et al. ⁶⁸ ). Morseth et al. ¹⁰⁹ showed an increase in blood pressure of normotensive female rats exposed daily to white noise (at 67–124 dB) for 8 h, during a two-week period. However, no change in blood pressure was noticed after lifelong (over 2 years) daily 10 h exposure to 85 or 105 dB broadband sound, which might be a consequence of habituation. ^67,68,110

Several studies have shown altered hormone levels in response to noise exposure. For example, increases in the levels of norepinephrine, cholesterol and corticosterone have been reported. The increase in stress hormone levels suggests a noise-induced activation of the hypothalamic–pituitary–adrenal axis (HPA axis), which might cause several problems related to abnormally elevated levels of circulating stress hormones. ⁶⁸ Armario et al. studied the effects of acute and chronic noise on serum levels of pituitary hormones in male Wistar rats. They found that acute noise stimuli increased serum levels of corticosterone, prolactin, and LH and decreased serum growth hormone (GH), while follicle-stimulating hormone was unaffected by this stressor. Chronic noise did not modify basal levels of any hormone studied; however, responsiveness of some hormones to the same stimuli was altered: reduced corticosterone, prolactin and GH responses to noise were observed after previous chronic exposure to this stimuli. ¹¹¹ Circulating testosterone levels were shown to be increased in Wistar rats that had been exposed to acute noise (1 h, 85 dB). ¹¹² A rapid doubling of rat plasma corticosterone levels (lasting for 2–4 h) was also shown as a result of exposure to banging of metal cages in an animal room. ¹¹³ More recently, Burow et al. showed a noise-intensity-dependent increase in the plasma adrenocorticotrophine hormone and corticosterone (HPA axis products) in Sprague-Dawley rats in response to the perceived threat of loud noise, with levels beginning to rise at approximately 85 dB (A). In this study, the rats were exposed to noise for a period of 30 min, ranging from 80 dB (A) to 110 dB (A) (SPL) in increments of 5 dB (A). c-fos mRNA induction (a marker of regional brain activity) was very low in the brains of the control and 80 dB (A) groups, but several brain regions displayed a noise-intensity-related induction of this marker. ¹¹⁴

The reproductive function of rats can also be affected by sounds. Zondek ¹¹⁵ showed that exposure of rats to ultrasounds of 50–80 kHz at 80–90 dB in the four days during the mating period reduced fertility by 73.2% and productivity by 84%. Exposure to 100 dB of 3–12 kHz for one minute during the four days of copulation reduced fertility by 70–80%.

Sleep disturbances induced by environmental noise have also been reported. Disturbances in sleep pattern were reported for pigmented Long-Evans rats after chronic exposure to environmental noise (similar to road traffic and railway noises: with a major octave band ranging from 20 to 3000 Hz) – composed of a background noise of 70 dB and several unpredictable noise events corresponding to a global intensity of 88 dB. A chronic exposure to this environmental noise (9 days) restricted continually amounts of slow wave sleep (SWS) and paradoxical sleep (or rapid eye movement) and fragmented these two sleep stages with no habituation effect. Rabat et al. ^116,117 also found evidence for the existence of subpopulations of rats that are either resistant or vulnerable to these deleterious effects of environmental noise on sleep and especially on SWS amounts, bouts number and bout duration.

An example of interference of sound with metabolism was reported by Friedman et al., ¹¹⁸ who found that plasmid lipid (triglyceride) levels were doubled in rats that had been administered an oil-enriched meal during continuous 102 dB white noise with an intermittent one second burst of 114 dB of 200 Hz.

Loud noise exposure has also been reported to cause cellular effects. For example, ultrastructural alterations in Wistar rats' myocardium ¹¹⁹ and adrenal glands ¹²⁰ have been shown in rats exposed to loud noise (100 dB) for 12 h. These ultrastructural changes were shown to involve mainly the mitochondria and endoplasmic reticulum. DNA damage in the above-mentioned organs was also found to be associated with this loud noise exposure. ^119–121 Myocardium alterations were concomitant with increased in situ noradrenaline levels and utilization. ¹¹⁹

Different levels of background noise were shown to influence learning and behaviour in rats.

In a study using outbred Wistar rats and testing several different sounds, Voipio showed that the behavioural responses depend on the type of sounds. Noise type of sounds caused fear reactions, such as startle, flight and freezing, even if at a low sound pressure level. Wave-type sounds induced movement or no response at low SPL, but more clear effects at high SPL. ⁶⁷

A decrease in locomotor activity of rats was shown in response to either natural 22 kHz calls or artificial 38 kHz signals, during and after ultrasound exposure. These Wistar rats were exposed to 5 min of non-stimulus sounds (background noise correspondent to tape recorder and tape noise) followed by 5 min of stimulus signal (further noise, 38 kHz signals or 22 kHz calls) and finally a further 5 min of background noise. The SPL of each of the ultrasonic signals was set to about 65 dB, similar to levels of 22 kHz calls recorded from defeated rats. Natural calls also decreased loudspeaker sniffing compared with the response if the speaker only emitted the background noise of the tape. Sales suggested this was probably because the animals have already experienced these 22 kHz calls in their social context, in their home cages, and might have learnt to respond to them with reduced activity. Calls emitted by other rats in the colony room, during group transportation or in the experimental room could affect the behaviour of experimental rats that can hear them. ¹²²

In open field behaviour, continuous white noise of 85 dB was shown to increase defecation and reduce both social activities and non-social activities (e.g. sniffing, grooming or crawling) of male Sprague-Dawley rats when compared with 50, 65 or 75 dB. As suggested by Weyers et al., ¹²³ the changes in social activities and defecation may be interpreted as anxiety reactions.

In a recent study, genetically-defined rats (DA inbred strain) learned a complex maze exposed to noise of moderate intensity (70 dB) or to a quiet environment (≤35 dB). Noise-exposed rats made fewer errors, explored less and finished their trials sooner. Previous studies have shown that noise of an intensity level as used in this study increased choline uptake in several brain regions including the prefrontal cortex and the hippocampus. ¹²⁴ Therefore the facilitating effects of noise might have been due to increased cholinergic activity. ¹²⁵ Thus, the acoustic environment is a factor that needs thorough control in studies with animal models of learning and memory.

Animals are able to adapt to sound. Outbred Wistar rats have been shown to adapt to sounds after repeated stimuli. During the exposure period to different sounds (16 days: 4 days with exposure to different sounds of 60 dB SPL; 2 days of pause; 4 days with exposure to different sounds of 80 dB SPL; 2 days of pause and 2 days with exposure to rat scream at 60 dB + 2 days with exposure to rat scream at 80 dB), the onset adaptation to less harmful sounds was short, but more harmful sounds caused intense response and habituation took longer. Furthermore, it was mentioned that the adaptation ‘memory’ was short and that similar type of sound caused similar response even on the next day. ⁶⁷

Noise in the animal facility

Several studies have been published showing the different sounds that can occur inside the animal facility where environmental noise is virtually unavoidable. The sources of sound can be technical devices (such as air-conditioners, air handlers, ventilated rack systems, electronics, video monitors, laboratory equipments and fire alarms); maintenance procedures done routinely (such as opening and closing doors, changing cages, cage washers, push carts, workers' speech and even rubbing cloth during movements and crinkling paper towels); and animals themselves by their movements (e.g. climbing and chewing on cages and accessories) and by their vocalizations (as mentioned previously). ^{42,68,98,99,126–130} For example, in the analysis of sounds associated with equipments used in an animal facility, Sales et al. recorded SPL of 56 dB (at 1 m distance) and 70 dB (at 0.1 m distance) in a high-frequency range for the two computer systems studied. The computer monitor appeared to be the source of most of the sound (10–100 kHz spectrum showing regular pronounced peaks at frequencies higher than 15 kHz). ¹²⁸ The constant frequency and continuous ultrasound produced by oscilloscopes (28 kHz with overall SPL of 48 dB at 1 m) and visual display units (producing a harmonic series of sound frequencies from 16 kHz up to 160 kHz, with overall SPL around 60 dB at 0.5 m) have been shown to reduce total activity of male rats in an open field arena. In this study, the sound output of taps running into a sink was measured, being characterized as a complex sound of continuous noise with broadband short bursts of sound within a frequency range up to 160 kHz with an overall maximal SPL of 95 dB. ¹²⁷ These are examples of (ultra)sounds that can be produced in the animal facility being apparently silent to humans but in the hearing range of laboratory animals, such as rats.

Sound levels are specially produced during animal care activities being variable between weekdays and weekends, and between day and night time. ^98,126 Milligan et al. recorded the acoustic environment of rooms housing rats over a 24 h period. In one of the facilities, there was a peak in sound levels about 90–100 dB in the lower frequency range during normal working hours. In the high-frequency range, the sound profiles were also increased during working hours with levels often reaching 70–85 dB. ¹²⁶

The sound exposure levels inside the rat cage and in the adjacent cage, while workers developed care procedures (e.g. pulling cage out of the rack, placing it onto a table and replacing the cage back into the rack, putting food into the food hopper), have been recently measured by Voipio et al. Hurried work with steel caused sound exposure levels exceeding 90 dB(R) when the cages were placed into the rack and about 80 dB(R) when pulling them out of the rack or placing them onto a table. With polycarbonate cages, the levels were 15 dB(R) lower. Unhurried calm working produced lower sound exposure levels than hurried working in many procedures (about 10–15 dB(R) differences in both cage materials). When the same procedures were performed in adjacent cages, the sound exposure levels were lower, but the differences were similar. When comparing rat and human hearing, H-weighted sound exposure levels (for human hearing sensitivity) were about 10–20 dB higher than R-weighted (for rat hearing sensitivity). ¹²⁹

The value of using masking noise, such as radios, to reduce adverse effects of impulsive noise has been discussed. Music treatment, produced by playing the Herbert Von Karajan Adagio compact disk through loudspeakers (less than 40 dB) from 09:00 h until 14:00 h daily during eight days, has been shown to decrease the metastaticity of WRC 256 cells in Sprague-Dawley male rats of both groups: submitted and non-submitted to auditory stress (exposing animals to the fire alarm bell (100 dB)). ¹³¹ Loudspeakers of most domestic radios have a limited sound output to below 16 kHz and, as shown by Sales et al., this would not cover the full frequency range of environmental noise which is within the range of most laboratory animals. Using radios in the animal house might therefore be a benefit to animal house staff, being indirectly a benefit to animals taken care of by those workers, ¹²⁸ but the effect of background noise, such as radio and different types of music, in masking the effects of ultrasound and loud noises needs further investigation.

Noise can be involved in several scientific studies, not as a controlled experimental variable but as an unintended environmental variable that can in some cases confound animal-based research results, ⁶⁸ interfering with animal welfare and stress.

A ‘silent fire alarm’ ¹³² was developed taking into consideration the difference in hearing abilities between humans and rodents, in order to fulfil the need for an effective alarm signal in laboratory animal facilities that did not stress the animals every time it was tested. The device developed produced pure tones alternating between 430 and 470 Hz, giving a sound level of 97 dBC at 450 mm – the energy in the alarm signal was below the optimal hearing range for mice and rats. Producing a noise ‘intensely irritating and disturbing’ for humans, this alarm did not awake rats and mice from sleep, and if already awake, rodents did not show a startle response, ear twitching or other indication of auditory disturbance when this ‘silent fire alarm’ was switched on. ¹³²

In conclusion, both environmental and communication sounds are present in animal facilities having a wide variety of effects in animals' physiology and behaviour, and consequently influencing animal welfare and animal-based research results. Environmental noises differ among different animal facilities and a careful planning should be made before construction of the animal facility, with the help of acoustic engineers, taking into account the differences in the sound perception between rodents and humans, ⁶⁹ in order to avoid stressful environmental sounds both for the animal and personnel. Procedures that may elicit both audible or ultrasound vocalizations should take place in a location that will not allow those sounds to reach other animals (unless that is part of the experiment). ⁴² Animal facility staff and researchers working at the animal facility should try to avoid uncontrolled ultrasounds (which might interfere with animals' communication sounds) and audible sound production. Measurements should be made in each animal facility in order to register existing environmental sounds and detect existing ultrasounds.

Cage cleaning (and transport)

Several authors have shown that cage cleaning and in-house transport promote an acute stress response on laboratory animals. Armario et al. reported that cage cleaning of adult male Sprague-Dawley rats, housed in groups of three, elevated serum concentrations of corticosterone ¹³³ and prolactin. ¹³⁴ These endocrine stress responses were measured 15 min after the onset of acute stress (when rats were bled by decapitation) and were dependent on the intensity of the stimulus. When rats were submitted to cage cleaning alone, the increases in plasma corticosterone and prolactin were significant but lower than when rats were submitted to cage cleaning followed by transport to a new room. Cage cleaning and transport to a noisy room (alarm bell of 85 dB) were more stressful than cage cleaning and transport to a quiet room. ^133,134 Barrett et al. reported a three-fold higher plasma corticosterone level in male Wistar rats bled after being transported to a laboratory compared with rats bled in their holding room. ^113,135 Moving adult male Han:Sprague rat cages from the racks to the floor was shown to result in a significant increase in plasma corticosterone levels within 5 min, reaching a peak at 15 min and returning back to baseline levels after 60 min. ¹³⁶ In this experiment, Gärtner et al. also showed a significant increase of serum levels of prolactin, thyroid-stimulating hormone (TSH) and triiodothyronine (T3) within 15 min after moving cages, and after 60 min the levels of these hormones in the serum of transported rats were still elevated. ¹³⁶ Regarding effects of moving cages on HR, Gärtner et al. reported an increase in HR from 340 to about 450 beats per minute (recorded telemetrically in unrestrained rats carrying a permanently implanted transmitter), tachycardia persisting for 10–15 min. Other effects on microcirculation, such as alterations of packed cell volume, haemoglobin and plasma protein, increased significantly within 2 min after moving cages and lasted about 10 min. ¹³⁶ Disturbances in metabolism, which were suggested to be a consequence of endocrine and microcirculatory alterations and the activation of the sympathetic–adrenal system, were also found after moving cages: plasma glucose increased significantly 3–8 min after moving cage; plasma lactate and pyruvate increased 2–3-fold above control levels 1–3 min after cage moving. These effects of cage moving on metabolism lasted only for 10 min. ¹³⁶

Activation of the sympathetic nervous system has also been associated with routine housing procedures such as cage cleaning. Significant increases in plasma noradrenalin and adrenalin were reported by Boer et al. immediately after gently lifting catheterized male Wistar rats (individually housed) from their home cages and placing them in a new cage (identical to the first cage but without bedding, food and water) for a period of 15 min before returning the animals back to their original cage. ¹³⁷ A significant increase in plasma corticosterone and glucose concentrations was also detected at 15 min after the placement of the rats into the new cage, returning to baseline levels at t = 45 min. The increased levels of catecholamines, corticosterone and glucose found in rats that have been handled and placed in a new cage were assumed to result from psychological activity in response to the change in environment (e.g. fear or emotional arousal), rather than being the result of physical activity or peripheral physiological changes induced by these environmental alteration. ¹³⁷

Even though using old anaesthetics that are not recommendable anymore, ¹³⁸ Tabata et al. found a significant rise in plasma glucose levels in mice following handling and transport of the cage to an adjoining room, with plasma glucose levels appearing to return to normal levels after about one hour. The same set of procedures, when performed on F344 and Sprague-Dawley rats (males and females), seemed to have small or no observable effect on levels of plasma glucose. ¹³⁹ The authors argue that rats might have acclimatized more easily to the handling procedure used than mice, which might have reduced the release of glucose into the bloodstream as a physiological stress response. ^135,139

Laboratory rats' behavioural responses to cage cleaning have also been addressed. The behaviour of male Sprague-Dawley rats in their home cages have been shown to vary with time of day and cleaning regime, with rats showing more activity on cleaning days increasing behaviours such as locomotion, grooming, digging and climbing and reducing sitting. ¹⁴⁰ Saibaba et al. interpreted this increased activity after cage changing as a sign of disturbance or a response to novelty, since the olfactory and possibly the visual and tactile environments were altered within the cleaned cage and could stimulate exploratory behaviours. Animals were also reported to tend to be more active in the morning periods than in the afternoon periods tested, which might be related to the arrival of staff and beginning of the working day with a general increase in noise levels. ¹⁴⁰

More recently, physiological responses to cage cleaning have been studied using radiotelemetry: increases in HR and blood pressure in both male and female rats have been reported in response to a variety of handling procedures. ^141–146 For example, studying the effects of routine cage cleaning on cardiovascular and behavioural parameters, Duke et al. observed a prompt increase in systolic, diastolic and mean arterial blood pressure (MAP), HR, and cage behaviour (such as movement, rearing and grooming) in adult male Sprague-Dawley rats when placed in clean cages. Elevations in HR and MAP in response to cage change had a duration of approximately 45–60 min. Rats witnessing this procedure did not show significant changes in HR or MAP. Manipulated animals also became aroused, showing increased awake, moving, rearing and grooming behaviours for at least 45 min whereas witnessing animals showed much less activation, and most returned to sleep by 15–30 min after cage change occurred. Adding a small amount of soiled bedding from the previous home cage to the new one did not modify increases of HR, MAP or behaviour, indicating that familiar olfactory cues did not counteract the novelty of the new cage. ¹⁴¹ However, in male mice it made a remarkable difference in aggressive encounters whether soiled bedding or the old paper nest was transferred into the new clean cage. ¹⁴⁷ Rats whose cage had not been changed for a period of two weeks presented an earlier onset of increased HR and MAP increased more rapidly in response to cage change, presenting a more prolonged cardiovascular response when compared with the response of rats whose cage has been changed weekly. ¹⁴¹ Grooming was less in the last 15 min of the observation period following cage change in the animals whose cage has been changed weekly compared with those whose cage had not been changed for a period of two weeks. Repeating these experiments in four consecutive weeks, researchers reported that cardiovascular and behavioural responses to the fourth change were not different from those observed in the first week, indicating the animals did not habituate to the procedure. ¹⁴¹

When measuring the cardiovascular response of adult male Sprague-Dawley rats to several common procedures (cage change, restraint and subcutaneous injection, restraint and tail-vein injection, exposure to odour of urine and faeces from stressed rats, and exposure to the odour of dried rat blood), by radiotelemetry, Sharp et al. ¹⁴⁵ also found significant increases in HR of rats housed individually, with one cage mate or with three cage mates, but HR in the four-to-a-cage rats decreased more rapidly, reaching baseline levels within 30 min. In response to routine cage change, rats housed individually or with one cage mate showed significant HR increases that were observed for 90–120 min after the movement to the new cage. Changes in MAP after cage cleaning showed the same patterns as HR, except in the case of rats housed four per cage in which MAP decreased significantly below baseline values from 90 to 180 min after the cage change. ¹⁴⁵ The decrease of HR and MAP below the baseline levels in the rats housed four per cage after routine cage change was suggested to result from the exposure to the odour of ammonia from soiled bedding that had accumulated in the cage during the four days since the cage was last changed. Manipulation of rats also affected their behaviour increasing the number of arousal behaviour such as moving, rearing and grooming, the increases being more prevalent in individually-housed than group-housed rats. ¹⁴⁵ When testing individually or group-housed adult female Sprague-Dawley rats, the same authors also found significant increases in HR following routine cage changing and for durations of 30–90 min before HR returned to baseline, suggesting the rats were stressed. ¹⁴³ In this study, transporting rats to another laboratory for subcutaneous injection was also shown to significantly increase HR (equivalently in all housing groups). Active behaviours of female rats in the home cage were increased for at least 30 min after the husbandry procedure. ¹⁴³ Group-housing frequently reduced the stress-like response. ^143,145 The physiological responses observed were similar during cage changing or simulated cage changing (where rats were returned to the original cage), suggesting the responses are caused by the physical manipulation rather than unfamiliar aspects of the new cage. ¹⁴³ No noteworthy differences were registered by Sharp et al. in the cardiovascular (HR and MAP) and behavioural responses induced by cage changing or witnessing cage changing, between females in different stages of the oestrus cycle (proestrus–oestrus and metestrus–diestrus), but the authors noticed that female rats showed more cardiovascular (stress-like) and arousal responses than did male rats in the previous study, suggesting an influence of gonadal hormone status. ¹⁴⁶ Female rats witnessing the cage change procedure also showed greater cardiovascular response than did male animals, and females showed noteworthy increases in response to room entry whereas males did not. Regarding home-cage behaviour, males showed two times more sleeping behaviour and much greater differences between manipulated and witnessing groups than females and no markedly effect of stage of the oestrus cycle was noticed in females. ¹⁴⁶

In another study by Sharp and collaborators, it was reported that witnessing routine procedures such as cage changing does not induce significant stress-like responses such as the ones observed in animals submitted to cage cleaning. Only Sprague-Dawley male rats housed alone or with only one cage mate had small increases in HR, MAP and home-cage behaviours while witnessing cage change. ¹⁴² In contrast, female Sprague-Dawley rats witnessing cage changes showed greater responses, but increases in HR and active home-cage behaviours were also reduced in females housed in groups than in single-housed females, suggesting that group-housing reduces stress or potential for stress. ¹⁴⁴

In 2006, a cross-laboratory study by Burn et al. ¹⁴⁸ reported that cage cleaning frequency had no clear impact on rat welfare. Male Sprague-Dawley and Wistar rats, housed in groups of four, were kept in four different animal units and their cages were cleaned twice-weekly, weekly or every two weeks, and contained either aspen woodchips or absorbent paper bedding. Among other parameters, aggression, injuries and general health, weight gain, handleability and in-cage ammonia concentration were monitored. Frequent cleaning decreased ammonia concentrations and handleability, and non-aggressive skirmishing was highest in weekly cleaned rats. However, no clear welfare benefit or harm was associated with any of the cage cleaning frequencies of these socially-housed male rats, as no differences in growth rates or general health have been found when comparing animals from frequently cleaned with less frequently cleaned cages. ¹⁴⁸ In agreement with previously mentioned studies (e.g. by Duke et al. ¹⁴¹ and by Sharp et al. ¹⁴⁵ ) when observing the behavioural activity of male Sprague-Dawley and Wistar rats immediately after cage cleaning, Burn et al. reported an increase in the general activity of rats, including walking, bedding manipulation and feeding, which were above baseline levels for the full 30 min observation period and during this period the number of rats resting did not return to precleaning levels. A marked increase in non-aggressive and play-like skirmishing to above precleaning levels was also noticed, but the effect was transient and after 15 min this skirmishing behaviour returned to below precleaning levels. Since cleaning frequency did not affect the magnitude of the postcleaning peak in skirmishing, Burn et al. ¹⁴⁹ suggested that the peak is neither caused by any relative change in the olfactory environment nor influenced by how habituated rats are to disturbance. In this study, rats were also observed to move more frequently to sheltered areas after cage cleaning, which might indicate they are trying to avoid light while undisturbed rats (in the day before cleaning) remained in their resting position kept since the dark period or it can indicate an exploratory behaviour motivated by the environment of the new cage. No behavioural evidence was found suggesting that the increased postcleaning activity was an acute stress response corresponding to a negative effect of cage cleaning on rat welfare. ¹⁴⁹ Chromodacryorrhoea (a dark red stress-related secretion produced by the Harderian gland in the orbit of the eye) was scored around the nose of rats after the behavioural observation period (i.e. 35–45 min after cage cleaning) and was found to be higher on the day before cage cleaning than after cleaning, suggesting that having soiled bedding was more stressful than the cleaning procedure itself, but since the scoring was made only after the observation period it might have been groomed away by rats before scoring (e.g. chromodacryorrhoea has been shown to be released between 16 and 30 min after acute stress induction by limb restraint). ^149,150

In-house transport effects

As previously mentioned, in-house transport of laboratory rats induces several physiological (endocrine, cardiovascular and metabolic) responses, such as increases in plasma corticosterone, prolactin, TSH and T3 ^{113,133,134,136} ; increase in HR; increases in haemoglobin and plasma protein; and increase in plasma glucose. ¹³⁶ Other reports can be found in the literature on behavioural and physiological changes resulting from in-house transport of laboratory rats, showing an acute stress response to the transportation procedure or to its witnessing. For example, the behaviour of non-transported Wistar Cpb:WU rats has been shown to be altered in the presence of rats that had been transported in their cages on top of a trolley pushed through the animal house for 2–3 min. ¹⁵¹ An open box with two adjacent small fields was used and each animal – one transported and one non-transported – was placed in each field and their behaviours were assessed: transported female rats showed significantly decreased sniffing and rearing and increased grooming (suggesting the induction of transportation stress), whereas non-transported rats displayed significantly increased sniffing and tended to urinate more frequently. Since there was no physical contact between transported and non-transported rats, de Laat et al. ¹⁵¹ suggested that communication might have occurred through sounds or odours from transportation-stressed rats affecting the behaviour of non-transported rats.

More recently, Dallmann et al. ¹⁵² have shown that moving the cage of group-housed male F344/Hw rats within the holding room or transferring the animals, in their cage, between holding and test rooms (through a noisy corridor with constant background noise caused by ventilation and pig holding rooms on the other side of the corridor) resulted in a significant increase in BT (T _b). When moving the cage inside the holding room, the stress-induced hyperthermia (SIH) corresponded to an elevation in rats' T _b of more than 0.5 ºC for the following 120 min. Transportation between holding room and test room was shown to increase rats' T _b over time lasting for at least 60 min. The fact that rats' SIH lasted for 120 min when transporting the cage from the rack to a table inside the holding room made Dallmann et al. suggest that, before starting an experimental procedure, habituation periods after in-house transport should be up to 120 min or even longer. ¹⁵²

Cage cleaning and contact with odours

Odours presented to laboratory animals during cage cleaning might represent a stressor and have an impact on animal wellbeing. Personnel performing the routines of cage changing and cleaning can be a source of odours, personal odours or odours from other cages or animal rooms. Rats might fear humans carrying scents from pets ⁶⁹ or from other animal rooms in the animal facility via fomites (clothing, hands), which is one important reason why all personnel working in the animal facility should wash their hands and change laboratory coats between rooms, ⁴² avoiding the transmission of odours (inter-species and between groups of animals of the same specie) and also avoiding contact of their clothes used outside the animal facility with the laboratory animals. Presenting predators' odours (e.g. cat fur/skin odours) to laboratory rats can elicit defensive behaviours, fear and anxiety (revised by Blanchard et al. ¹⁵³ ). Natural human odours or perfume and deodorants' odours associated to humans might also be stressful to the laboratory rats. For example, odorants and smells (volatile chemicals), often being present in perfumed products, can influence HPA activity and immune responsiveness ¹⁵⁴ or be anxiolytic to rats. ¹⁵⁵ Even though the effect of human odours (natural or artificial ones) in rats needs further investigation, humans should avoid bringing predators' smells, perfumes and deodorants' odours into the animal facility where they are in contact with laboratory animals, during routine husbandry procedures (such as cage cleaning) or experimental procedures.

Cage cleaning and ventilation

Another aspect that affects cage cleaning and animal welfare is ventilation. The air movement in the cage is related to other important environmental factors such as temperature (affecting the thermoregulatory capacity of animals) and relative humidity. Ventilation and air renewal at cage level affects its microenvironmental air quality in terms of microorganism concentrations, dust particles and noxious gases (such as ammonia and carbon dioxide). ¹⁰ To establish the frequency of cage cleaning in an animal facility, all these factors have to be considered because of their effects on the animals. For example, the prevalence of pneumonia was shown to increase with ammonia levels inside the cage (from 25 to 250 ppm) and pathological changes were found in the respiratory tract of rats exposed to soiled bedding with 100–200 ppm NH₃ levels, which can be found in rat cages (without filtered tops) after a week without cleaning. ^156,157 Most rapid ammonia production occurs under conditions of high humidity ¹⁵⁷ and the type of bedding also has an influence. ¹⁵⁸ Ammonia and carbon dioxide levels in individually ventilated cages (IVC) were studied by Silverman et al. during a seven-day period without cage cleaning. In this study, ammonia levels reached values ≥ 500 ppm after three days and intracage carbon dioxide increased rapidly until values ≥ 10,000 ppm. ¹⁵⁸ In another study analysing different ventilation rates and cage changing frequencies and their impact on C57BL/6J mice house in ventilated cages, Reeb-Whitaker et al. ¹⁵⁹ concluded that cage changes once every 14 days and ventilation rates of 60 air changes per hour (at which ammonium levels were about 50 ppm in trio-mated mice) provide optimum conditions for animal health and practical husbandry and had no adverse health effects. Further studies are needed in order to have more information on the welfare consequences of increased ammonia concentrations and there are no guidelines for the maximum ammonia concentration to which rodents can be exposed to. Based on guidelines for human exposure and veterinary literature, Silverman et al. suggested an intracage concentration of 50 ppm should lead to cage cleaning of mice housed in IVC or disposable static cages. ¹⁵⁸ Different IVC systems might also provide different cage microenvironments and each case has to be investigated.

As mentioned, constant air renewal is needed and adequate air changes in the rats' close environment are essential to control their microenvironmental temperature, humidity and air quality. ¹⁰ Animal facilities usually have between 15 and 20 air changes per hour in animal rooms, as indicated in Appendix A of Council of Europe Convention ETS 123. ² IVC systems enhance the ventilation inside the cages to higher rates of air changes per hour allowing facilities to reduce cage change frequency, but these systems constitute a relatively novel source of potential discomfort. For humans, air speed levels exceeding 0.2 m/s is considered draughty and this is also generally agreed to be an upper limit for rodents. ¹⁶⁰ Higher intracage ventilation rate could induce chronic stress and heat loss due to the draught. It has been shown in preference tests that rats choose cages with lower than 80 air changes per hour. ¹⁶¹ In respect of mice, their physiology and behaviour were not affected when there were fewer than 80 air changes per hour, the air inlet came from the top and nesting material was provided. Thus, the location of the air supply in the cage (from the wall or from the top), the ventilation rate and the presence of nesting material are important when considering the impact of IVC housing on mice wellbeing. ¹⁶²

In conclusion, common husbandry procedures were indeed shown to induce stress-like responses in laboratory rodents. As mentioned above, cage cleaning and in-house transport have a considerable impact on physiological and behavioural responses in rats, characteristic of an acute stress response, but a clear impact on rats' welfare has not been proven yet. Researchers should be careful with the animal facility cage cleaning schedules when planning studies. Proper ventilation systems should provide a good ventilation control in the animal facility and consequently at the cage level. To establish the cage cleaning frequencies, animal facility personnel should have the possibility of making measurements of the intracage microenvironment (e.g. humidity, temperature, noxious gas levels) and these data should also be available for researchers. Care should be taken in dismissing husbandry procedures as non-stressful just because they are routine. These common procedures might influence experimental data if not taken into account or controlled for. For example, when measuring peripheral endocrine responses to stressors, researchers should perform blood sampling quickly (e.g. within 100 s of first touching the animals' cage) to avoid influences of the stress response to the animal/cage manipulation itself on experimental results ¹³⁶ ; in studies that require determination of basal cardiovascular parameters, Sharp et al. ¹⁴³ recommended that data should not be obtained for at least 2 h after common husbandry procedures. By applying radiotelemetry technique in mice, Meijer showed that responses of HR and BT paralleled corticosterone responses to various routine procedures (e.g. different methods of restraint, injections by different routes, cage cleaning). Given the acute stress responses of laboratory animals to routine procedures, Meijer ¹⁶³ stated that basal values of HR, BT or corticosterone should not be assessed directly after the animals have been subjected to such procedures – a recovery period of at least one or two hours was recommended in order to improve experimental procedures and to obtain reliable data for basal measurements.

Other stressors during cage cleaning might be the contact with humans and the odours they carry. When handling the animals for cage cleaning, all personnel should avoid the use of clothes that were in contact with rats' natural predators, the use of deodorants and perfumes, and use the appropriate protection clothing and procedures against odours spreading.