All work was done at the Department of Clinical Sciences of Companion Animals, Faculty of Veterinary Medicine, Utrecht University, The Netherlands
Effects of Aging on Brainstem Responses to Toneburst Auditory Stimuli: A Cross-Sectional and Longitudinal Study in Dogs
Version of Record online: 28 JUN 2008
Copyright © 2008 by the American College of Veterinary Internal Medicine
Journal of Veterinary Internal Medicine
Volume 22, Issue 4, pages 937–945, July–August 2008
How to Cite
Ter Haar, G., Venker-van Haagen, A.J., Van Den Brom, W.E., Van Sluijs, F.J. and Smoorenburg, G.F. (2008), Effects of Aging on Brainstem Responses to Toneburst Auditory Stimuli: A Cross-Sectional and Longitudinal Study in Dogs. Journal of Veterinary Internal Medicine, 22: 937–945. doi: 10.1111/j.1939-1676.2008.0126.x
- Issue online: 4 JUL 2008
- Version of Record online: 28 JUN 2008
- Submitted September 11, 2007; Revised November 13, 2007; Accepted April 17, 2008.
- Age-related hearing loss;
- Frequency-specific evoked responses;
Background: It is assumed that the hearing of dogs becomes impaired with advancing age, but little is known about the prevalence and electrophysiologic characteristics of presbycusis in this species.
Hypothesis: As in humans, hearing in dogs becomes impaired with aging across the entire frequency range, but primarily in the high-frequency area. This change can be assessed quantitatively by brainstem-evoked response audiometry (BERA).
Animals: Three groups of 10 mixed-breed dogs with similar body weights but different mean ages were used. At the start of the study, the mean age was 1.9 years (range, 0.9–3.4) in group I, 5.7 years (3.5–7) in group II, and 12.7 years (11–14) in group III.
Methods: In a cross-sectional study, the BERA audiograms obtained with toneburst stimuli were compared among the 3 groups. In a longitudinal study, changes in auditory thresholds of group II dogs were followed for 7 years.
Results: Thresholds were significantly higher in group III than in groups I and II at all frequencies tested, and higher in group II than in group I at 4 kHz. The audiograms in group II indicated a progressive increase in thresholds associated with aging starting around 8–10 years of age and most pronounced in the middle- to high-frequency region (8–32 kHz).
Conclusions and Clinical Importance: Age-related hearing loss in these dogs started around 8–10 years of age and encompassed the entire frequency range, but started and progressed most rapidly in the middle- to high-frequency area. Its progression can be followed by BERA with frequency-specific stimulation.
In humans, age-related hearing loss (ARHL), or presbycusis, is one of the most prevalent chronic health conditions among the elderly and the most common form of sensorineural hearing loss encountered in industrialized nations.1–6
Although it is difficult to separate the effects of noise injury on hearing ability from those of aging, chronic noise exposure leads to substantially increased hearing thresholds in the elderly, and noise exposure at an early age may trigger progressive hearing loss later in life.7,8 In addition, ARHL in mice has been found to have a genetic basis, and 10 loci that promote ARHL thus far have been identified.6,7 Evidence for a genetic effect on the inheritance of presbycusis in women and a mixed, genetically acquired cause in men is supported by data from Gates et al.9 In addition to noise and heredity, systemic degenerative changes also appear to contribute to the development of ARHL. The process of aging is associated with many molecular, biochemical, and physiologic changes, including increases in mitochondrial DNA damage and reduction in mitochondrial function.10,11 Reactive oxygen species-dependent destruction of sensory epithelia has been shown to play an important role in presbycusis.10,12 In the broadest sense, the term presbycusis therefore refers to the cumulative effects of heredity, disease, noise, ototoxic agents, and probably other environmental and dietary factors superimposed upon those of the aging process itself, although it is sometimes used in relation to the effect of aging only.1,3,6
Many studies have documented age-related hearing impairment in humans.13–18 These fall into 2 categories, cross-sectional studies and longitudinal studies. Cross-sectional studies, describing differences among age groups, have shown that pure tone hearing thresholds increase with age, particularly at high frequencies.1–3,5 Longitudinal studies measure changes with age within individuals, where each subject can act as his or her own control. This type of study provides a better description of the course of changes with age and identifies intrinsic interindividual differences. There have been few reports with longitudinal design in humans.2,13,16–18 Substantial reduction in hearing capacity appears to occur from the age of 60 years onward and begins at the high frequencies (6–16 kHz), but gradually encompasses the entire frequency range.1,2,13–18 Between the ages of 70 and 80 years, the reduction of hearing amounts to 1–2 dB/y, depending on the specific frequency tested.2,13,14,16,17
Previous reported studies of hearing loss in dogs have concerned the hearing capacity of puppies, have been cross-sectional in nature, or have employed click stimulation in brainstem-evoked response audiometry (BERA).19–24 Because hearing loss can be partial and starts at the higher frequencies in humans, its accurate description in dogs requires determination of audiograms over the entire frequency range. The technique of BERA using frequency-specific stimulation of the canine cochlea over a wide audiometric range, from 1 to 32 kHz, was described previously and was used for this study.25 Recently, other techniques for determination of high-frequency hearing loss in dogs have been reported, of which the use of auditory steady-state evoked potentials (ASSEP) seems to have the most potential.23,24
Hearing loss in humans affects the individual's psychosocial situation and, if left untreated, contributes to social isolation, depression, and loss of self-esteem.1,2 Although it is known that dogs also develop hearing loss with advancing age, little is known about its prevalence, electrophysiologic characteristics, or psychosocial effects in this species.1,19–21
The aim of this study was to test the hypothesis that dogs, similar to humans, develop hearing impairment across the entire frequency range with increasing age, first and most noticeably in the high-frequency area, and that this impairment can be quantitatively assessed by BERA by toneburst stimulation. Results of both a cross-sectional and a longitudinal study in dogs by BERA with frequency-specific stimulation of the cochlea over a wide audiometric range (1–32 kHz) are presented.
Materials and Methods
The protocol of this study was approved by the Animal Experimentation Committee of the Faculty of Veterinary Medicine, Utrecht University, the Netherlands.
This investigation consisted of a cross-sectional study and a longitudinal study. The cross-sectional study compared the frequency-specific BERA thresholds (audiograms) of 3 groups of dogs with substantially different mean ages, designated groups I (young), II (middle-aged), and III (geriatric). At the start of the study, the mean age in group I was 1.9 years (median, 1.8; range, 0.9–3.4; n = 10), in group II it was 5.7 years (median, 5.5; range, 3.5–7.0; n = 10), and in group III it was 12.7 years (median, 13.3; range, 11.0–14.0; n = 10). Smaller breeds of dogs live longer than do larger breeds and are considered to become geriatric at a later age.26 To enable comparisons to be made between middle-aged and geriatric dogs, only dogs from the same weight category using the Goldston and colleagues classification were used in groups II and III.26 They categorized dogs, according to their weight, as being small (0–9 kg), medium (10–23 kg), large (24–40 kg), or very large (>40 kg) and considered these categories to become geriatric at 11.5, 10.9, 8.9, and 7.5 years, respectively.26 Using this classification, all dogs in groups II and III were medium-sized and all in group III were geriatric (>10.9 years).
Group I consisted of 10 privately owned female dogs of different breeds having a mean weight of 20.8 kg (median, 21.5 kg; range, 9.0–30.6) referred to the Faculty of Veterinary Medicine, Utrecht University for neutering. Their hearing was assessed during recovery from surgery. Group II consisted of 10 clinically healthy dogs housed in the kennels of the department. Their mean body weight was 17.8 kg (median, 18.6; range, 12.5–21.3 kg) at the start of the study. Seven were mixed-breed littermates and 3 were Beagles. Five were intact males, 1 was a castrated male, and 4 were intact females. The initial audiograms of these dogs were reported previously.25 For the longitudinal study, frequency-specific hearing thresholds were determined in this group once yearly or once every other year for 7 years, from a mean age of 5.7 years to a mean age of 12.3 years. Group III consisted of 10 clinically healthy elderly dogs maintained in the kennels of the department from an early age onward. They included 6 mixed-breed intact females and 4 Beagles, of which 3 were intact females and 1 was an intact male. Their mean body weight was 15.8 kg (median, 15.3; range, 13.2–21.5).
None of the dogs had any clinical signs of neurologic or otic disease, and otoscopic examination under anesthesia verified normal external ear canals and normal tympanic membranes in all ears. The privately owned dogs had no known history of ototoxic drug administration or hearing impairment and had not been exposed to excessive noise. The 20 dogs from the department kennel had no history of ototoxic drug administration but were exposed to the noise of barking dogs in the kennel. The sound pressure level of this noise measured with a sound level metera with A-weighted filter settings was 82 dB (dBA SPL; A-weighting according to the frequency dependent response of the [human] ear [IEC 179b and ISO 1999c]) with short peaks of 104 dB SPL during feeding. Because these dogs were not housed separately or trained to respond individually to commands from the technicians, there was no simple means of determining their hearing ability other than that they all exhibited what was considered to be normal behavior.
In groups II and III, a light plane of anesthesia was induced with medetomidined (100 μg/kg IV), followed by propofole (1 mg/kg IV) within 5 minutes and was maintained with these drugs during auditory testing, as described previously.25 At the end of the procedure, anesthesia was antagonized with atipamezolef (250 μg/kg IV). Rectal temperature was maintained at 36.5–37.5 °C by means of a circulating water heating pad placed under the dog. In group I, anesthesia was induced in the same way with medetomidine and propofol, after which the dogs were intubated and surgical anesthesia was maintained by inhalation of isofluraneg (delivered in a 1:1 mixture of oxygen and air). After completion of ovariectomy and closure of the abdominal wound, isoflurane administration was stopped and dogs were allowed to recover to a light plane of anesthesia, similar to that in groups II and III, for auditory testing, after which anesthesia was antagonized. Although different anesthesia protocols were used, the effect of anesthesia on the brainstem auditory-evoked responses (BAER) is generally considered to be negligible and the protocols used in this study have not been reported to affect the BAER.25
Auditory testing with determination of hearing thresholds by means of BERA was performed as described previously.25 Briefly, with the dog in sternal recumbency, 3 recording needle electrodesh were inserted SC, 1 at the base of the pinna of each ear and the third as the reference electrode over the occipital protuberance on the midline. The nonstimulated ear was plugged with an adjustable wax plug.i The electrodes were attached to an amplifierj with a total gain of 20,000–50,000 and a bandpass filter of 3–300 Hz. Click stimuli (CS) were generatedk as rectangular waves of 0.2 ms duration. Toneburst stimuli (TS) of 1, 2, 4, 8, 12, 16, 24, and 32 kHz were generated by software, transmitted to a digital-to-analog converterDACl, and delivered to the ear via a high-frequency bandpass speaker (tweeterm) with a frequency response up to 38–40 kHz. The tweeter was connected to the ear via a flexible, well-fitting transmission tube inserted deep in the ear canal. Stimuli alternated between rarefaction and condensation with a nominal repetition frequency of 10 Hz. The intensity output of the tweeter was measured at the end of the flexible tube with a sound pressure metern to determine the stimulus sound pressure levels in dB SPL. The measurements were made with a linear filter setting of 1–70 kHz. The output of the tweeter was checked with both a waveform analyzero and a microphonep from which the electrical output was displayed on an oscilloscope to visually confirm the shape and duration of the stimulus and the number of peaks. All tonebursts were confirmed to have a center lobe at the aimed frequency. Side lobes of the nominal frequency were present, but their power was at most a few percent of that of the central lobe. All amplified response signals were fed to an analog-to-digital converterq interfaced to a personal computer. Additional details on signal acquisition, averaging, and analysis controlled by dedicated software were reported previously.25
For each ear, hearing was assessed by delivering CS, beginning at 80 dB SPL and decreasing in intensity in steps of 10 dB until the threshold was reached. Two recordings were made for each set of stimulus variables to check reproducibility. The entire procedure then was repeated, frequency-by-frequency, using TS of 1, 2, 4, 8, 12, 16, 24, and 32 kHz to determine the threshold at each frequency. Threshold was defined as the intensity 5 dB above the 1st decreasing level at which no visually recognizable brainstem wave V was elicited. Wave V was identified as the wave preceding the deepest negative trough in the tracing more than 3.5 ms after stimulation.25 Wave V was chosen to determine threshold because its amplitude is usually the least affected by aging and the last to disappear with decreasing stimulus levels.1,23,27 When no response to stimulation was observed at 80 dB SPL, the intensity of the stimulus was increased in steps of 10 dB to a maximum of 100 dB. If still no response was seen at 100 dB, then 100 dB was arbitrarily assigned as the threshold value. In group I, both ears were tested once in a single session. In group III, both ears were tested in 2 separate sessions 3–12 weeks apart. In group II, both ears were tested in 2 separate sessions 3–12 weeks apart and then this procedure was repeated at intervals of 12–24 months until (a) there was a marked increase in the pure tone threshold, or (b) the dog reached the median age of group III, or (c) the dog was removed from the study for unrelated reasons (eg, malignancy).
For the cross-sectional study, a linear mixed effects modelr,s was applied with a normal distribution for threshold. Two random effects associated with each dog were included: a random intercept and a random frequency level effect. The threshold was the dependent variable and the independent grouping variables were the ear, frequency level, age group, and the 2-way interactions among the 3 independent variables. A variance model was used, which allows a different variance per frequency level. The final model included the effects of frequency level, ear, the interaction between ear and age group, and the interaction between age group and frequency level.
A paired Student's t-test was used to determine whether significant differences existed in group III between threshold increases (compared with groups I and II) in the low-frequency area (1–4 kHz) and the high-frequency area (8–32 kHz).
For the longitudinal study, the initial model was the same as for the cross-sectional study. Three random effects associated with each dog were included: a random intercept, a random effect for moment of observation (age), and a random frequency level effect. A variance model was used, which allows a different variance per moment of observation (age). The final model included frequency level, ear, and the interaction between moment of observation (age) and frequency level.
In both models, the restricted maximum likelihood method was used to estimate the covariance parameters. The maximum likelihood method was used for estimating the fixed effects. The Akaike's Information Criterion was used to select the best model.28 The models were fitted by the statistical program R, version 2.5.1.rP≤ .05 was considered significant.
The brainstem-evoked responses after click and toneburst stimulation resulted in the characteristic 5–7 positive-peak pattern at 80, 90, or 100 dB SPL in all but one of the dogs tested.19,25,29–33 In the exception, a dog aged 13.9 years in group III, there was no response to 32 kHz toneburst stimulation at 100 dB SPL but there were measurable thresholds at all other frequencies tested. In all dogs, as the intensity of the stimulus decreased the amplitude of the evoked response decreased, while peak latency increased until the threshold was reached (see Fig 1). There were marked differences in thresholds at different frequencies not only in the same animal, as reported previously,25 but also among the 3 age groups and within group II during the longitudinal study.
The composite audiogram for the 3 groups in the cross-sectional study with the mean threshold values is shown in Figure 2. The difference in thresholds between groups I and II was not significant except at 4 kHz, where the threshold was significantly higher in group II (P < .05). The cross-sectional study identified markedly higher thresholds in group III than in groups I and II, the difference being significant at all frequencies tested (P < .05). The highest absolute thresholds were in the middle- to high-frequency region (8–32 kHz), and the increase in thresholds was significantly greater (P < .01) for this region than for the low-frequency region (1–4 kHz) in group III compared with both groups I and II. None of the differences in threshold between the left and right ear was significant.
Individual audiograms were constructed for all 10 dogs from group II in the longitudinal study, 6 of which are shown in Figure 3. Dogs II-H, II-J, and II-K were still alive at the end of this study, the other 7 having been euthanized during the study, most often because of malignancy.
Dogs II-A, II-B, and II-C were tested 3 times during the period between the mean ages of 6 and 9 years (see Fig 3A for the audiogram of dog II-A). No increase in threshold was significant in these 3 dogs, nor were any of the differences in threshold between the left and right ear significant.
The audiograms of the other 7 dogs disclosed a progressive increase in hearing thresholds with aging, starting at 8–10 years of age. The effect was most pronounced at the middle to high frequencies (8–32 kHz). However, there were considerable differences in severity and rate of hearing loss among these dogs (see Fig 3A and B for the audiograms of dogs II-D, II-E, II-G, II-J, and II-K).
The average decline in hearing at all tested frequencies in group II between the mean ages of approximately 6 and 12 years is shown in Figure 4. All dogs of group II were still alive at a mean age of 10 years (n = 20, both ears tested), but only 7 dogs (n = 14) were still alive at a mean age of 12 years. Three dogs were still alive at the end of the study, with a mean age of 14 years, but this number was too small to be included in these calculations. Individual thresholds increased gradually over the years starting around 8–10 years of age (Fig 4). Statistical analysis of the average increase in thresholds identified a significantly higher threshold at a mean age of 12 years than at a mean age of 6 years for 8, 12, 16, 24, and 32 kHz (P < .05).
Presbycusis, or ARHL, is the major type of hearing loss and the predominant neurodegenerative disease of aging in humans.34 There have been many reports about this increasingly common form of sensorineural hearing loss in humans, and its incidence, prevalence, auditory brainstem response characteristics, and pure tone thresholds have been described in detail in experimental, cross-sectional, and longitudinal studies and in review papers.1–18,27,34,35 Presbycusis most likely reflects the cumulative effects of heredity, disease, noise, and ototoxic agents superimposed on those of the aging process itself.1,3,6 Especially, the effects of noise injury are difficult to separate from those of aging, and there are obvious interactions between them. Not only is considerable hearing preservation demonstrated in individuals raised in a relatively noise-free environment, but also animal models have shown the existence of 2 windows of increased susceptibility to noise exposure, early in life (adolescence to early adulthood) and late in life, with subsequent increases in hearing thresholds as a result of degeneration of sensory epithelium later in life.7 Studies in humans have reached similar conclusions.8 In a recent study, it was found that early noise exposure in mice can trigger a progressive neuronal loss later in life.7
The decrease in function of the cochlea as a result of aging leads to hearing loss, which can be assessed with both behavioral studies and electrophysiologic techniques. Hearing thresholds usually are measured behaviorally in humans, especially in adults.1,13 Even though this is also possible in dogs after appropriate training, as has been shown by Heffner36 among others, it is very impractical in the clinical setting and thus objective assessment of hearing requires electrophysiologic measurements. Thresholds measured behaviorally are highly correlated with those measured electrophysiologically in humans, although the latter tend to be somewhat higher, especially in the elderly, which can lead to a slight underestimation of hearing ability.27 The difference depends on the stimulus frequency and ranges from several dB at high frequencies to as much as 15–20 dB at lower frequencies.37,38 In considering this difference in thresholds, it should be noted that frequencies of 8 kHz and above are high for human hearing, whereas dogs can hear much higher frequencies. Both the electrophysiologically defined thresholds reported in our previous study and those found in this study in the group I dogs are in close agreement with behaviorally measured thresholds in the original report by Heffner on hearing in large and small dogs.25,36 The lowest thresholds in the latter study were found to be between 2 and 16 kHz. For 3 of the 4 dogs reported in Heffner's study, the thresholds at 16 kHz were equal to or lower than those at 2 kHz. Whereas the lowest absolute thresholds were found in the 8 kHz area compared with the 12 kHz area in our study, threshold differences between 8 and 16 kHz in the Heffner study are only between 0 and 5 dB in 3 of the 4 dogs. Unfortunately, 12 kHz was not examined separately, but we can conclude that thresholds measured behaviorally in dogs are highly correlated with those measured electrophysiologically.
BERA has been the most commonly used electrophysiologic method for assessing hearing in dogs, and the most common stimulus has been a click with a broad spectral energy distribution. Although it is known that the threshold for click-evoked potentials can give a rough idea of the overall threshold for hearing, it cannot provide accurate information about hearing capacity over the entire audible frequency range.39,40 Because hearing loss in presbycusis primarily begins in and mostly affects the higher frequencies in humans, frequency-specific thresholds should be determined in dogs as well. This can be achieved with either toneburst stimulation or by masking clicks with high-pass noise or noise with a spectral gap.22,23,38,41–44 The applicability of techniques with masking to increase frequency selectivity was considered, but these techniques were found to be unreliable by Gorga and Worthington45 because of the tuning curve of the high-frequency neurons. Threshold measurements in dogs have been reported using toneburst stimulation, but only in puppies or in dogs without hearing impairment.22,25,46 A major concern in using tonebursts is the spread of energy to frequencies other than the nominal center frequency, which is also known as spectral splatter.22,47 Thus, the BAER threshold obtained with tonebursts is not completely determined by the response of the neurons at the center frequency but also that of the side lobe frequencies. Recruitment of lower and higher frequencies by the side lobes could have altered the responses at threshold in our normal hearing dogs and hence could have made the results less frequency specific. However, as mentioned before, the results are in close agreement with those obtained with behavioral methods. In patients with hearing loss, these side lobes could lead to an underestimation of hearing loss at the nominal frequency of the toneburst.47 However, this seems to be particularly true for hearing loss limited to specific frequencies, such as noise-induced hearing loss, where surrounding frequencies usually are unaffected. In presbycusis, where pathology is not limited to a small area or one specific frequency, this problem is probably less relevant. Theoretically, hearing loss at certain frequencies could have been greater than reported in this study, and this phenomenon could have contributed to the flatness of the slope of the audiogram of the aged dogs (group III) in this study.
A promising new technique for frequency-specific threshold measurement in dogs was reported recently by ASSEP.24 Although the technique appears to be a valid method for obtaining frequency-specific thresholds in dogs, the authors state that work remains to be done on the low variability of interindividual thresholds, and frequencies of 16 kHz or higher were not tested. Although still inferior to behavioral testing, because of problems with spectral splatter in the tonal stimuli and the aforementioned discrepancy between behaviorally and electrophysiologically obtained thresholds, the most direct approach to obtaining frequency-specific thresholds in the dogs at the beginning of our study was frequency-specific TS.25,39
To the best of our knowledge, ours is the 1st report on presbycusis in dogs documented with audiograms by BERA. There have been reports of decreased hearing in aged dogs documented by BAER but with click stimulation or only 1 stimulus level. No thresholds have been reported in aged dogs with impaired hearing by means of either click or toneburst stimulation. Knowles reported increased latency and decreased amplitude of responses in 4 dogs with reduced hearing and the absence of recognizable peaks at 84 dB SPL stimulus intensity in 5 dogs.19,20 In a study of age-related changes in the cochlea and cochlear nuclei, Shimada et al21 reported that in 13 dogs older than 10 years the responses to click stimulation at 90 dB SPL were diminished or absent.
Cross-sectional studies in humans have shown that pure tone hearing thresholds increase with age, particularly in the high frequencies.1–3,5,13 Although cross-sectional studies can describe differences among age groups, the observed differences are confounded by cohort differences, making it difficult to match the demographic and clinical features of young and older subjects.2,13 In addition, the potential for bias must be considered in these studies. In our study, information bias was minimized, and all materials and methods as well as data interpreters were identical from the beginning to the end of the study. Also, selection bias appears to have been small in this study, because all dogs of groups II and III were chosen at random from the same pool and were used throughout the entire study, and their housing, diet, and other environmental factors were unchanged.
The most important conclusion of our cross-sectional study is that auditory thresholds at all frequencies tested were significantly higher in elderly than in young and middle-aged dogs. Although differences were significant at all frequencies, the most dramatic increase in thresholds was seen at middle to high frequencies (8–32 kHz), which is similar to hearing loss in human presbycusis.1,2,13–18 Considering the audiogram in dogs, which encompasses a much wider range of audible frequencies than in humans, especially high frequencies, the greatest loss of hearing capacity in elderly dogs occurred in the area of the audiogram with the lowest initial thresholds in dB SPL. The cochleas of all dogs of group III have been examined histologically, and the findings are compatible with sensorineural hearing loss. These findings will be reported separately.
The number of dogs studied was relatively small for a cross-sectional study, but the individual differences within each group were small, which lends credence to the findings. Relative to the lifespan of dogs, those in group III were of advanced age and were all considered to be geriatric. Most otogerontologic studies in humans available for comparison have reported hearing loss occurring between the ages of 65 and 80 years.13–18 Presbycusis is reported to be slight in normal individuals at the age of 60 years, but thereafter a significant reduction in hearing capacity occurs, beginning at the high frequencies (6–16 kHz).1,2,13–18,34 Between the ages of 70 and 80 years, the reduction of hearing is obvious and generally amounts to 1–2 dB/y, depending on the specific frequency tested.2,13,14,16,17 In addition, the amount and rate of decrease in hearing capacity also depend on gender and noise exposure as well as the level of hearing capacity at the start of the investigation.2,17,27 Presbycusis is more prevalent and more severe in men than in women and the rate of hearing deterioration in men is more than twice that experienced by women.3,48 These authors state that decreases in hearing sensitivity may be detected at all frequencies in men by age 30, whereas the onset is later in women. Once the impairment is clinically detectable, women tend to have better hearing than men at frequencies above 1,000 Hz, whereas men have better hearing at lower frequencies.3,48 In view of the small number of male dogs in this study, sex differences could not be examined.
The contribution of kennel noise exposure to the observed increase in auditory thresholds is unknown, but probably played a role in both groups II and III. Noise-induced hearing loss in humans typically occurs in the 2–6 kHz range, with initial changes at 6 kHz and the greatest increase in threshold at 4 kHz.2,16,49 This suggests that the increased threshold at 4 kHz in group II could have been the result of noise damage. Ambient noise measured in the kennel of groups II and III, mainly produced by the barking dogs, was not high enough to produce such alterations in humans.49 However, the influence of frequently occurring high levels of ambient noises on hearing thresholds in dogs is unknown. Evidence from mice and other species (including cats, hamsters, and guinea pigs) has indicated 2 windows of increased susceptibility to noise exposure with implications for thresholds and assumed ARHL later in life: early in life (up to about 4 months) and late in life.7 The dogs of groups II and III had been housed lifelong in kennels with sound-reflecting walls, and noise-induced hearing loss could also account for some of the differences between groups I and II. In light of the other thresholds in group II, especially when compared with group I, the 4 kHz threshold stands out and could conceivably have been because of unexpected focal interference or technical errors that were not recognized during the measurements. The increase in the 4 kHz threshold in the cross-sectional study in group II over that in group I is unlikely to be the result of presbycusis.
Although thresholds could still be determined in all of the dogs in group III, most were dramatically increased in comparison with those in the other 2 groups, implying that presbycusis was present in a very advanced stage. To study the progression of presbycusis before the end-stage is reached, 1 or 2 additional age groups would have been desirable in the cross-sectional study and would have enabled more accurate comparisons with human thresholds. Nonetheless, the present study clearly demonstrates the hearing loss associated with aging in dogs.
A longitudinal design overcomes many of the problems of a cross-sectional study design by measuring changes with age within individuals, where each subject can act as its own control, thus providing a better description of the course of changes with age and identifying intrinsic interindividual differences. Although environmental effects also differ among subjects in longitudinal studies, because it takes many years to obtain results and it is difficult to follow subjects over long periods of time, longitudinal studies are the preferred method for studying presbycusis.2,13,16–18
This study indicated that thresholds in group II gradually increased with advancing age from the age of 8 years onward in all dogs in which there was sufficient follow-up. There were differences in audiograms among dogs but not between ears of the same dog. The audiograms of group II are difficult to compare because the number of audiograms and the intervals between them were not similar for all dogs. Also, there was a 4-year difference in age between the youngest and oldest dogs in the group, and not all dogs could be followed to old age.
There were enough measurements, however, to enable comparison of thresholds in all dogs at the same age, from a mean age of 6 years to a mean age of 12 years. Interesting conclusions can be drawn from the changes in auditory thresholds tested at octave frequencies from 1 to 32 kHz during these years. Thresholds at a mean age of 12 years were significantly higher than those at a mean age of 6 years for 8, 12, 16, 24, and 32 kHz (P < .05), demonstrating that presbycusis starts in the middle- to high-frequency area in dogs.
The average increase between 10 and 12 years was lowest (around 10 dB) at 1, 2, and 4 kHz. Increases in thresholds were significantly greater at higher frequencies, ranging from 15 dB at 8 kHz to more than 22 dB at 12, 16, 24, and 32 kHz. This difference between low and high frequencies also is observed in longitudinal studies in presbycusis in humans. In 2005, Lee et al2 reported the average rate of change in thresholds in humans to be 0.7 dB/y at 0.25 kHz, 1.2 dB at 8 kHz, and 1.23 dB/y at 12 kHz. Enrietto et al13 also reported an average yearly increase of 0.8 dB in the 0.5 kHz threshold and 1.7 dB in the 8 kHz threshold. The threshold increases that we observed in dogs are 10 times greater than those reported in humans. This may in part be explained by our use of 10 dB steps in lowering stimulus intensity, but using steps of 5 dB would double the anesthesia time and would be very impractical for clinical use in dogs. Another hypothesis for the larger increase in thresholds in aging in dogs could be the dog's much shorter lifespan, so that any age-related deterioration might occur at a more rapid rate than in humans.
Although our observations and conclusions are derived from a small group of medium-sized dogs and therefore not necessarily applicable to all dogs, we conclude that presbycusis does exist in dogs, as in humans. The impairment can be demonstrated and its progression can be followed by means of BERA with frequency-specific stimulation of the cochlea. Our findings indicate that increases in auditory thresholds occur from a mean age of 8–10 years onward, eventually encompass the entire frequency range, and occur initially and are greatest at middle to high frequencies (8–32 kHz).
aBrüel and Kjaer microphone (type 4135) and measuring amplifier (type 2610), Naerum, Denmark
bIEC 179: Precision sound level meters, 1973
cISO 1999: Determination of occupational noise exposure and estimation of noise-induced hearing impairment, 1990
dDomitor, Pfizer Animal Health, Capelle a/d IJssel, the Netherlands
eRapinovet, Mallinckrodt Veterinary, Houten, the Netherlands
fAntisedan, Pfizer Animal Health
gIsoflurane, Abbott Laboratories Ltd, Maidenhead, Berkshire, United Kingdom
h13L61, 35 mm, 0.7 mm diameter, Dantec Medical, Scovlunde, Denmark
iOhropax, OHROPAX GmbH, Wehrheim, Germany
jAB601G, Nihon Kohden Co, Tokyo, Japan
kGrass stimulator S88, Astro-Med Inc, Grass Instrument Division, West Warwick, RI
lBiophysical Laboratory, Department of Clinical Sciences of Companion Animals, Utrecht University, the Netherlands
mDHT9, 8Ω, frequency range: 1-40 kHz, Visaton, Haan, Germany
nBrüel & Kjaer, type 2231
oStanford Research Systems, SR760 spectrum analyzer, Sunnyvale, CA
pBrüel & Kjaer, 1/4 inch, type 4136, bandpass (−6 dB; 20–100 kHz)
q14 bit, PCL 816, America Adventech Corp, Sunnyvale, CA
rR Development Core Team (2007). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org
sJose Pinheiro, Douglas Bates, Saikat DebRoy and Deepayan Sarkar, the R Core team (2007). nlme: Linear and Nonlinear Mixed Effects Models. R package version 3.1–86
The authors are grateful for the assistance of Harry de Groot, the Department of Anesthesiology and the supporting technicians, Dr J.C.M. Vernooij for his help with the statistical analysis of the data, and Dr B.E. Belshaw for editing.
- 3Presbycusis. In: CanalisRF, LambertPR, eds. The Ear: Comprehensive Otology. Philadelphia, PA: Lippincott Williams & Wilkins; 2000:545–557., ,
- 26Introduction and overview of geriatrics. In: GoldstonRT, HoskinsJD, eds. Geriatrics & Gerontology of the Dog and Cat. Philadelphia, PA: WB Saunders Company; 1995:1–9.
- 28In all likelihood: Statistical Modelling and Inference Using Likelihood (Oxford Science Publications). Oxford: Oxford University Press; 2001.
- 34Age-related hearing loss: The status of Schuknecht's typology. Curr Opin Otolaryngol Head Neck Surg 2004;12:439–443.
- 36Hearing in large and small dogs: Absolute thresholds and size of the tympanic membrane. Behav Neurosci 1983;97:310–318.
- 39Frequency specificity in evoked potential audiometry. In: JacobsenJT, ed. The Auditory Brainstem Response. San Diego, CA: College-Hill Pres; 1985:147–177., , , et al.
- 40Clinical evaluation of the auditory evoked brain stem response. Auris-Nasus-Larynx 1975;2:97–105., , , et al.
- 45Some issues relevant to the measurement of frequency-specific auditory brainstem responses. Sem Hear 1983;4:353–362.,
- 47Brainstem auditory-evoked potentials. In: SilmanS, SilvermanCA, eds. Auditory Diagnosis, Principles and Applications. San Diego, CA: Academic Press Inc; 1991:249–297.,