The study was presented at the 2007 American College of Veterinary Internal Medicine Forum, Seattle, WA.
Corresponding author: Dr Jacques Penderis, Institute of Comparative Medicine, Division of Companion Animal Sciences, Faculty of Veterinary Medicine, University of Glasgow, Glasgow, G61 1QH,UK; e-mail: firstname.lastname@example.org.
Background: The brainstem auditory-evoked response (BAER) is currently the standard evaluation method of hearing in dogs. In asymmetrical hearing loss in human patients, simultaneous presentation of masking noise to the nontest ear is routinely performed during BAER to eliminate the crossover effect.
Hypothesis: The crossover effect occurs during canine BAER, and masking noise of 20 decibels (dB) below click stimulus intensity is sufficient to abolish this effect.
Animals: Fifty-six Dalmatian puppies with confirmed unilateral deafness.
Methods: The BAER was elicited with 80 and 100 dB normalized hearing level (dBnHL) stimulus intensity in the deaf ear. The 100 dBnHL stimulus was repeated while simultaneously applying 80 dBnHL white masking noise to the nontest ear.
Results: Ten dogs were excluded because of BAER trace baseline fluctuation. In the remaining 46 dogs, 8 dogs had no waveforms, but 38 dogs had an identifiable wave-V in the deaf ear BAER at 80 dBnHL intensity stimulus. At 100 dBnHL intensity stimulus, all but 1 dog had a discernible wave-V in the deaf ear BAER. The deaf ear BAER waveforms were abolished by white masking noise at 80 dBnHL in the nontest ear in all dogs.
Conclusions and Clinical Importance: Abolition of BAER wave-V in the deaf ear by white masking noise in the nontest ear suggests that this wave is caused by the crossover effect. β distribution indicates 95% confidence that white masking noise, at 20 dB below click stimulus intensity, would abolish this crossover effect in over 90% of the dogs. This supports using masking noise in the nontest ear during canine BAER.
The brainstem auditory-evoked response (BAER) is one of the most commonly used tests for assessing auditory function in dogs and is also useful as a site-of-lesion diagnostic tool.1 Measurement of BAERs is noninvasive, widely available, and a simple method. It allows assessment of conductive auditory pathways in the outer and middle ears, sensory auditory structures in the inner ear, cranial nerve VIII, auditory parts of the brainstem, and higher neural structures involved in auditory perception.1
Canine deafness is identified with increased frequency because of heightened awareness about this condition among owners and breeders and also among clinicians.2 The most commonly diagnosed forms of deafness are congenital sensorineural deafness; conductive deafness associated with otitis externa, otitis media, or both; and later-onset sensorineural deafness (associated with chronic otitis interna, otitis media, or both; ototoxicity; noise trauma; or presbycusis in older animals).2 Currently, BAER is the main validated technique for assessment of congenital sensorineural deafness in dogs, and deafness has been reported in over 80 breeds of dogs.2 Deafness in the Dalmatian breed is associated with the extreme-white piebald gene2 and has been histologically described as cochleo-saccular degeneration.3,4 Deafness results from concurrent degeneration of the stria vascularis and organ of Corti and is accompanied by collapse of the saccule.3,4
Although masking noise in the nontest ear is routinely performed in auditory assessment of humans and is sometimes used in auditory assessment of dogs, its clinical relevance has not been validated in dogs. Furthermore, there is a poor understanding of what constitutes an effective level of masking noise in dogs. The rationale for using masking noise is to eliminate the crossover effect. This effect can occur during testing of a deaf ear when the test stimulus crosses over to the nontest ear, evoking contralateral activity.1 It is important that this crossover effect is abolished to avoid incorrectly labeling a deaf ear as having a hearing threshold, albeit with reduced hearing sensitivity. The aim of this study therefore was to establish whether the crossover effect occurs during BAER of dogs and to assess whether masking noise of 20 decibels (dB) below click stimulus intensity is sufficient to abolish this effect.
Materials and Methods
BAER was performed in 56 Dalmatian puppies with previously diagnosed unilateral deafness. The hearing status of the ears was assumed to be normal based on the recording of the typical BAER waveforms, with at least 4 positive peaks in response to 80 dB normalized hearing-level (dBnHL) click stimuli with simultaneous presentation of white masking noise to the contralateral ear. Deafness was diagnosed if an isoelectric line was obtained in response to the same protocol. The Dalmatian breed was selected for the study because, based on histological studies, complete deafness could be assumed to be caused by the complete degeneration of the organ of Corti in affected animals.3,4
All dogs included in the study were client-owned pet dogs presented to a hearing assessment clinic at the Animal Health Trust. The group size was based on available case numbers and the previous demonstration that masking of the nontest ear appeared to abolish equivocal abnormalities in a number of puppies with unilateral hearing loss in 1 study.5 The dogs were not sedated for the procedure and were between 5 and 10 weeks of age when the BAER was performed, by which age click BAER wave-V characteristics have reached adult values.6 A Medelec Sapphire 2ME 2-channel electrodiagnostic systema was used, with amplifier filter settings comprising a low filter of 100 Hz and a high filter of 3 kHz. Disposable stainless-steel subdermal needle electrodes were used (impedance <2 kilo ohms), with the active electrode placed on the vertex, the reference electrode rostral to the tragus of the test ear, and the ground electrode on the dorsal midline of the neck. The BAER was elicited with 512 unilaterally applied 0.1 ms click stimuli (polarity set as rarefaction at 80 dB) generated by headphones (unshielded TDH49P audiometric headphones) manually held against the opening of the external ear canals. The initial stimulus intensity in the deaf ear was 80 dBnHL and subsequently increased to 100 dBnHL. The 100 dBnHL stimulus was then repeated with simultaneous application of 80 dBnHL broad-bandwidth white masking noise (noise signal containing all audible frequencies at equal intensity) to the nontest (normal) ear. The BAER of the unaffected ear was then obtained with a stimulus intensity of 80 dBnHL for confirmation of a normal waveform. Any dogs demonstrating irregular fluctuation of the BAER trace baseline because of the excessive movement of the puppies during execution of the test were excluded because of the potential for this to interfere with the recognition of BAER trace waveforms. Statistical analysis was performed by Minitab 15 Statistical Software.b Descriptive statistics for wave-V amplitude (measured from the peak of the wave to the lowest point of the following negative trough) and wave-V peak latency of the deaf ear at both 80 and 100 dBnHL and of the normal ear at 80 dBnHL were calculated. Mean amplitude ± SD and mean latency ± SD of the waves V are reported. A paired t-test was used to compare the amplitude and latency of the wave-V in the BAER of the deaf ear at 80 and 100 dBnHL stimulus intensity to the amplitude and latency of the wave-V in the BAER of the contralateral (normal) ear.
Ten dogs were excluded from analysis because of the fluctuation of the BAER trace baseline. These fluctuations were considered to be a result of excessive movement by the unsedated puppies during acquisition of the BAER. In the remaining 46 dogs, 8 dogs had no distinguishable BAER waveforms in the deaf ear, but in 38 (83%) dogs a wave-V was recognized in the BAER of the deaf ear at the 80 dBnHL intensity stimulus (Fig 1). The mean wave-V amplitude in these puppies was 0.84 ± 0.31 μV, and the mean latency was 5.16 ± 0.32 ms. At the 100 dBnHL intensity stimulus, all but 1 (98%) dog had a discernible wave-V in the BAER of the deaf ear. The mean wave-V amplitude was 1.23 ± 0.45 μV, and the mean latency was 4.64 ± 0.32 ms. After the use of white masking noise at 80 dBnHL in the nontest ear, no BAER waveforms could be identified in the affected ears. A normal BAER waveform was present in the unaffected ear in all puppies at a stimulus intensity of 80 dBnHL, confirming normal hearing status in this ear. In the normal ear, the mean wave-V amplitude was 2.17 ± 0.62 μV and the mean latency was 3.67±0.25 ms. Significant differences were evident in the amplitude and latency of wave-V in the BAER of the deaf ear at both the 80 and 100 dBnHL stimulus intensities, as compared with the amplitude and latency of wave-V in the BAER of the contralateral (normal) ear (P<.0001 for all) (Fig 2). β distribution indicated 95% confidence that white masking noise, at 20 dB below click stimulus intensity, would abolish this crossover effect in over 90% of the dogs.
In human audiology, masking of the opposite ear is recommended when there are substantial sensitivity differences between the left and right ears.7,8 If masking is needed, then a broad bandwidth of noise is considered to be the best masker when using a click stimulus, as the click spectrum encompasses a broad frequency range.7 There does not appear to be a requirement for the use of masking in human patients with normal hearing, as its use in these individuals does not appear to substantially affect either the latency or the amplitude of the BAER, even when the intensity of the masking noise is equal to the stimulus intensity.8
In most of the puppies assessed in our study, a wave-V was identified in the deaf ear after the 80 dBnHL stimulus. When the stimulus intensity was increased to 100 dBnHL, all but 1 of the animals had a wave-V, with a slight increase in amplitude and a decrease in latency in the deaf ear compared with the response to the previous stimulus. The increase in amplitude and decrease in latency with increasing stimulus intensity are consistent with what would be expected if the BAER were performed in normal dogs.1 Of the 7 wave responses recognized in the BAER, wave-V is used clinically to assess the threshold of hearing. The BAER wave response identified in the deaf ear in our study if no white masking noise is used was considered to be a wave-V originating from the contralateral ear caused by the crossover effect. In the crossover effect, the click stimulus directed at the test ear stimulates the cochlea of the nontest ear, either as a result of noise leakage to the nontest ear or the click stimulus causing vibration of the skull and thereby directly stimulating the contralateral cochlea. Complete deafness can be assumed in the Dalmatian to be caused by complete degeneration of the organ of Corti, with no residual hearing expected in animals affected by this form of congenital sensorineural deafness. Based on the characteristics of the wave identified, we also can assume it to be a wave different from that of the N3 potential. These vertex-negative potentials have been identified in human patients with neuro-epithelial deafness and in animals with suspected cochleosaccular deafness.9
It is important that this crossover effect is abolished by masking of the nontest ear to avoid incorrectly labeling the deaf ear as having a hearing threshold, albeit with reduced hearing sensitivity. The amplitude and latency of the wave-V in the BAER of the deaf ear, at both the 80 and 100 dBnHL stimulus intensities, are significantly different from the amplitude and latency of the wave-V in the BAER of the contralateral (normal) ear. Direct comparison of these results is imprecise because all of the wave-Vs reported originate from the unaffected ear but are recorded either with the reference electrode ipsilateral (when recording from the normal ear) or contralateral (when recording from the affected ear). Changing the side of the reference electrode mainly affects wave-I, but it also decreases the amplitude of wave-V.10 Nonetheless, the difference in latency may allow partial differentiation between the crossover effect and the presence of a hearing threshold. However, the overlap between the wave-V latency range of the normal and the deaf ear and the limitation of our assessment to puppies with no fluctuation of the BAER trace baseline means that in practice this difference in latency is unlikely to allow accurate distinction between the crossover effect and a hearing threshold in the deaf ear if masking of the nontest ear is not performed.
Insert earphones would probably have been advantageous as they reduce the stimulus artifact without influencing the acoustic signal and do not inadvertently collapse the ear canal, as may occur when the stimulus is generated by headphones.11 The influence of the use of insert earphones was not evaluated but it is likely that this method would have decreased the stimulus artifact (by decreasing the vibration of the skull, decreasing noise leakage, or both) and therefore diminish the crossover effect.
The use of contralateral masking noise levels of 30 dB below the click level delivered to the test ear has been recommended and did appear to be sufficient to abolish equivocal abnormalities in a number of puppies with unilateral hearing loss in 1 study.1,5 Other authors have suggested the use of higher levels (−20 dB) as a precautionary measure.12 On the basis of these studies, we elected to use a masking noise level of 20 dB below the click stimulus, as this still would be within the range of intensity that would be well tolerated by the dog, but would allow us to be confident that the level of white noise was sufficient to abolish the crossover effect.8 The use of a higher or a lower intensity of white noise masking was not evaluated because this study was carried out on client-owned pet dogs and this would have required longer procedure times and sedation of several of the puppies.
The abolition of the wave-V recognized in the BAER of the affected ear by the application to the nontest (normal) ear of white masking noise at 20 dB below the intensity of the click stimulus in all dogs in our study supports the use of masking noise. The β distribution of the results of our study indicates that we would expect with 95% confidence that the use of white masking noise, at 20 dB below the intensity of the click stimulus, would be effective in abolishing this crossover effect in over 90% of the dogs.