Vestibular evoked potential (VsEP) before and after noise exposure. Short-latency linear VsEP recordings before (3 repetitions above) and immediately after (4 repetitions below) exposure to 113-dB sound pressure level broadband noise for 1 hour. The intensity of the linear acceleration stimulus was 3g. Note the reduction in the amplitude of the first wave.
Percentage of change in wave I amplitude. Mean ± SD percent change in peak-to-peak amplitude of wave I of the vestibular evoked potential as a function of time before (0) and at several time intervals after the cessation of the noise. Also shown is the mean percentage of amplitude change of wave I of the major control group C (noise exposed but nonfenestrated).
Biron A, Freeman S, Sichel J, Sohmer H. The Effect of Noise Exposure in the Presence of Canal Fenestration on the Amplitude of Short-Latency Vestibular Evoked Potentials. Arch Otolaryngol Head Neck Surg. 2002;128(5):544-548. doi:10.1001/archotol.128.5.544
Exposure to high-intensity noise causes little, if any, reduction in vestibular function in normal animals as shown by short-latency vestibular evoked potentials (VsEPs).
To investigate the effect of noise exposure on VsEPs following fenestration of the horizontal semicircular canal.
Design and Methods
Psammomys obesus (fat sand rat) underwent labyrinthectomy in 1 ear, while the lateral semicircular canal in the other ear was fenestrated. Control VsEPs to linear acceleration (approximately 3g; rise time, approximately 1-2 milliseconds) were recorded immediately after the operation. The experimental group animals were then subjected to loud white noise (113-dB sound pressure level) for 1 hour. Immediately after the noise exposure in the experimental group animals, VsEPs were once more recorded.
The VsEPs in the experimental group animals were significantly reduced immediately following the noise exposure, while there was no change in the recordings from the control group animals (fenestrated but not noise exposed; noise exposed but not fenestrated), even though the noise exposure induced a mean 47-dB threshold elevation of the auditory brainstem response.
The presence of the fenestration caused the vestibular end organs to become vulnerable to noise exposure. The fenestration may create a pathway enabling pressure release through the vestibular end organs during noise exposure, thus increasing the possibility of damage to the vestibular end organs. This did not occur in the intact, nonfenestrated animals.
THE MAMMALIAN inner ear contains an auditory end organ sensitive to sound and several vestibular end organs that are sensitive to angular and linear acceleration. Outer perilymph channels and inner endolymph channels interconnect all end organs. Changes in perilymphatic pressure in response to a sound stimulus are transmitted throughout these channels to all parts of the inner ear. The basic sensory unit is identical in all parts of the inner ear: stereocilia-bearing hair cell, a primary sensory neuron, and the synapse between them. The transduction mechanism is also identical in all parts of the inner ear: mechanical changes in the fluids cause bending of the stereocilia. Bending in one direction causes depolarization of the hair cell and excitation of the relevant neural pathway, whereas bending in the opposite direction causes hyperpolarization and inhibition. This structural and functional similarity between the vestibular and auditory end organs has led to studies designed to test the possibility that sound stimuli can effect the vestibular end organs, as originally reported by Tullio1 (known as the Tullio phenomenon). For example, intense sound stimuli (130-172–dB sound pressure level [SPL]) have been shown to induce reflex eye movements in guinea pigs and monkeys.2,3 In humans, 125-dB SPL sounds have caused visual field shifts.3 Histological changes in the saccule and additional vestibular end organs following intense sounds (136-163–dB SPL) have also been reported.4 Also, primary vestibular neurons in cats have been shown to respond to sound stimuli of intensities 80- to 90-dB SPL.5,6 Clinical studies on healthy humans who had been exposed to noise and humans during noise exposure reported only subclinical vestibular findings.7,8 However, 17 patients in whom intense sounds induced vertigo and eye movements have recently been described; in each of these patients there was an absence of bone (dehiscence) over a portion of the superior semicircular canal (SCC).9
Over the past few years, a new objective technique for assessing the vestibular system has been developed: short-latency vestibular evoked potentials (VsEP), which is a noninvasive method for measuring potentials in response to acceleration impulses, similar to the technique used to record the auditory nerve brainstem response (ABR). These potentials were measured first in animals in response to impulses of angular10 and linear acceleration11 and later in humans in response to angular acceleration.12 The first wave (WI) of the ABR and VsEP represents the compound action potential of those auditory and vestibular nerve fibers synchronously activated by the respective stimuli and hence reflects end organ and primary auditory and vestibular nerve function.
These VsEPs have been used in our laboratory to objectively study the effect of noise on vestibular function in several controlled experimental paradigms in laboratory animals.13 The animals were exposed to noise intensities of 113-dB SPL for periods ranging from seconds to 3 weeks. There was either no effect of these exposures on the VsEPs or small effects of short duration with recovery.13 Thus, in response to several intensities and durations of noise, little if any vestibular changes were seen.
It has been suggested that this absence of an effect of high-intensity noise on vestibular function in normal ears could be because the round window, which is located in the cochlea, serves as a pressure release in the perilymphatic channel. Therefore, the sound pressures induced in the cochlear perilymph by stapes footplate vibration are preferentially transmitted to the perilymphatic channels of the cochlea and not to the vestibular channels. Hence, most of the damage is seen in the cochlea. Accordingly, induction of a fenestration (fistula) in one of the vestibular channels may enable the perilymphatic pressure wave to be transmitted through this alternative pathway to the vestibular end organs and make the vestibular end organs also noise sensitive. In fact, it has been shown in pigeons that fenestration of a semicircular canal makes vestibular neurons more sensitive to sound.14 In addition, single vestibular neurons of deaf mice began to respond to sound after fenestration.15
The present study was designed to clarify the mechanism that protects the vestibular organs from damage caused by exposure to noise. For that purpose, short-latency VsEPs to linear acceleration were recorded before and after exposure to high-intensity noise in animals with and without a fenestration in a semicircular canal.
The present study was performed on fat sand rats (Psammomys obesus). Using the sand rat has a major advantage in such an experiment owing to its unique middle and inner ear anatomy.16 The temporal bone of the sand rat contains an unusually large bulla. Most parts of the inner ear bulge into the bulla cavity and are easily accessible for delicate surgical procedures, including fenestration of one of the semicircular canals. The experiments were conducted in accordance with the guidelines published by the Hebrew University–Hadassah Medical School Animal Care and Use Committee.
All of the sand rats were anesthetized by an intraperitoneal injection of 35 mg/kg of pentobarbital sodium (Nembutal; Abbott Laboratories, Abbott Park, Ill) solution. Additional anesthesia was administered as required. A labyrinthectomy was performed on the nontested ear. This was done so that the recordings would originate from the tested ear only to facilitate interpretation of the results. Labyrinthectomy also saves the need for masking the other ear during ABR testing. The VsEP and ABR were recorded at several stages of the experiments.
Details of the stimulus and recording techniques have been previously reported.11 The stimulator comprised a solenoid coupled to a sliding device to which the animal's head, placed in a head holder (held between its nasal bone and hard palate), was attached. The device pulled the head linearly forward (utricle stimulus) with a linear acceleration of 3g and a rise time of 1 to 2 milliseconds (ms). Each stimulus caused a 50-µm movement of the head and was presented at a rate of 2 per second. After each stimulus, the head was returned to its original position at a lower acceleration by springs. The electrical activity was recorded with subcutaneous E2 platinum needle electrodes (Grass Instrument Division, Astro-Med Inc, West Warwick, RI) at the vertex referred to the left ear, with the ground electrode in the right ear. Conventional evoked potential equipment (Microshev 4000 evoked response system; Microshev Ltd, Efrath, Israel) was used. The recorded electrical activity in a 12.7-ms poststimulus window was amplified, filtered (150-3000 Hz), and the responses to the 128 stimulus repetitions were averaged. The VsEP responses were displayed such that a positive potential recorded by the vertex electrode appeared as an upward deflected wave. Throughout all VsEP (and ABR) recordings, body temperature was measured and maintained at 37°C to 38°C.
An earphone was placed 1 cm from the tested ear, through which a 120-dB peak equivalent (pe) SPL click was presented at a rate of 20.6 stimuli per second. Alternating polarity was used to cancel electrical artifacts. The intensity level was lowered in 5-dB steps until a threshold was reached (ie, the minimal stimulus intensity at which ABR waves could be identified). At least 2 readings were taken at every intensity level. Recordings were made with the same electrodes, sites, filter, average settings, and equipment used for the VsEP recording.
The following response parameters were evaluated: (1) peak-to-peak amplitude and peak latency of WI of the VsEP to 3g stimulation, (2) peak-to-peak amplitude and peak latency of WI of the ABR to 120-dB pe SPL, and (3) ABR threshold. An intensity of 120-dB pe SPL was used to assess ABR WI because after exposure to the noise, ABR often could not be recorded in response to lower-intensity stimuli.
A Grason-Stadler 455C noise generator (Grason-Stadler Inc, Milford, NH) connected to an auditory amplifier and a speaker were used to generate broadband noise. Spectral measurements were made using a precision integrating sound level meter type 2218 and third octave filter type 1625 (Bruel & Kjær, Copenhagen, Denmark). The level of noise intensity used was 113-dB SPL, with a peak at 2 kHz and 14 dB lower in intensity at 250 Hz and 26 dB lower at 125 Hz.
Following left labyrinthectomy, the lateral SCC of the right ear was exposed in all animals. The 26 animals studied were divided into the following 4 groups (Table 1):
Group A (7 animals): the effects of noise exposure on the VsEPs in the fenestrated ear were examined
Group B (5 animals): the ear not exposed to noise was fenestrated to assess possible effects of the fenestration itself on vestibular function
Group C (5 animals): the effects of noise exposure on the VsEPs in the intact (nonfenestrated) ear were examined
Group D (9 animals): the noise-exposed ear in which ABR was recorded was fenestrated to evaluate the effects of the noise exposure on the ABR of the fenestrated ear.
In groups A, B, and D, the bony SCC was fenestrated (size, 2 mm) with a small drill, taking care not to cause endolymphatic leakage; the SCC was not fenestrated in group C. The ABR was then recorded in the fenestrated groups B and D before the noise exposure, and the VsEP was recorded in groups A, B, and C before exposure. The animals in groups A, C, and D were then exposed to the 113-dB SPL noise for 1 hour (group B was not exposed to the noise). Immediately after the cessation of the noise exposure, recordings were again made of the VsEP in groups A, B, and C and ABR in groups B and D. In the groups in which both VsEP and ABR were recorded, ABR was recorded first. The VsEP and ABR were also recorded in the nonexposed group B 1 hour after the first recording. The VsEP recordings in the fenestrated group A were repeated every 1.25 minutes (approximately) for 30 minutes and again 1 hour after the end of the noise exposure. The VsEPs in the nonfenestrated group C were recorded every 1.25 minutes for 10 minutes. The sand rats were then given an intraperitoneal injection of a lethal dose of pentobarbital. Postmortem recordings were carried out to prove that the measurements were of a biological source and to rule out possible electromagnetic or electromechanical artifacts. The animals were under anesthesia during the entire experiment, from the time of the unilateral labyrinthectomy until after the postexposure recordings had been made, after which they were given the lethal dose. The maximal duration of anesthesia was approximately 2.75 hours. Thus, there was no need to remove the animal from the head holder and stimulating system between recordings, making the entire procedure well controlled.
We used 1-way analysis of variance (ANOVA) to compare amplitude and latency of VsEP recordings as a function of time in groups A and C. When a significant (defined as P<.05) correlation was found, post hoc t tests were used to compare the 2 specific points in time. The nonparametric paired Wilcoxon signed rank test was used to compare changes in groups B and D.
The ABR response parameters before and after noise exposure were compared to confirm that the level of noise exposure used (intensity and duration) was capable of causing auditory (ABR) threshold shifts. The ABR threshold and the amplitude and latency of ABR WI were compared before and after noise exposure (1 hour between). The ABR threshold was significantly elevated in group D after exposure to noise by 47 ± 17 dB (P = .01). The actual average threshold change was higher than this due to the inability to measure thresholds when they were greater than 120-dB pe SPL. At 120-dB pe SPL, the latency of ABR WI was significantly prolonged (P = .02) after noise exposure and amplitude was reduced, although nonsignificantly. No change was found in the ABR threshold, WI amplitude, or latency (120-dB pe SPL) in group B, recorded 1 hour apart. Thus, the noise exposure used in this study induces a clear and significant threshold elevation of ABR, and the presence of a fenestration for 1 hour does not affect cochlear (ABR) function.
This same noise exposure in the absence of a fenestration, but following exposure of the SCC (sham operation, group C), did not have any effect on the VsEP (WI before, 1.63 ± 0.38 µV; WI after, 1.74 ± 0.70 µV; 1-way ANOVA, F9,30 = 0.35; P = .95). In addition, the presence of a fenestration, but without noise exposure (group B), did not effect the VsEP (P = .1), confirming that the fenestration itself did not have a deleterious effect on vestibular function.
The most interesting result of this study is shown in Figure 1, in which the amplitude of WI of the VsEP can be seen to be reduced following noise exposure in the presence of a fenestration. The mean ± SD peak-to-peak WI amplitude reduction in group A immediately following the noise exposure was 60% ± 17.32% (P<.001) compared with the preexposure amplitude in each animal. This reduction was still present 1 hour later (1-way ANOVA, F22,130 = 1.966; P = .01; Figure 2), with a tendency toward partial, though nonsignificant, recovery. Latency did not change (1-way ANOVA, F22,130 = 0.555; P = .95).
Furthermore, 1 animal showed no change in VsEP recordings after its SCC was supposedly fenestrated and exposed to noise. In this animal, a postmortem examination revealed that the fenestration had not been properly carried out (the bony SCC was not successfully fenestrated). This animal was therefore excluded from the study.
It is generally accepted that the vestibular end organs are less vulnerable to high-intensity noise exposure than the cochlea in the intact animal. Several explanations have been suggested for these findings. First, high-intensity noise has been shown to cause lower blood flow in the stria vascularis17 and to lower the amplitude of the endocochlear potential.18,19 It is likely that this decreased strial blood flow is responsible for the lower endocochlear potential. This would depress cochlear transduction. However, because the endolymphatic spaces in the vestibular end organs do not have such a positive potential,20 transduction in the vestibular end organs is not dependent on such a large potential difference across the hair cell. This would render the vestibular end organs less sensitive to noise.
Second, it has been thought that the vestibular end organs are sensitive mainly to low frequencies and would not be affected by noise that does not contain much energy at lower frequencies. However, the noise exposure used in the present study was broadband, including low frequencies (eg, at 250 Hz, only 14 dB lower from the peak at 2 kHz). In addition, it has been shown that the vestibular organs can be excited by higher frequency stimulation because the vestibular stimuli used to elicit the short-latency VsEPs have high rise times (1-2 ms), ie, comprising high-frequency components (up to about 250 Hz).
And third, as explained above, the anatomical structure of the inner ear preferentially directs acoustic energy entering through the oval window to the cochlea due to the presence of the round window. Hence, most of the damage would be seen in the cochlea.
The present study was designed to test this last hypothesis, and the results have clearly shown that SCC fenestration followed by noise exposure caused a significant reduction in VsEP WI amplitude. Thus, it seems that the absence of a vestibular "round window" serves to limit stimulation of the vestibular end organs by acoustic stimuli. Vestibular end organ depression by the levels of noise used in this study (113-dB SPL for 1 hour) became apparent only in the presence of a fenestration and were not noted following exposure to the same noise without fenestrating the SCC. There was a slight nonsignificant tendency toward partial recovery following the initial sharp reduction of the amplitude of WI of the VsEP recorded immediately after the cessation of the noise exposure (Figure 2). In nonfenestrated Sabra rats exposed to 113-dB SPL broadband noise for 2.5 minutes, a VsEP amplitude reduction was also seen (10%-25%), but it was of short duration; recovery was apparent within 5 minutes.13 In the present experiment, it seems that the fenestration of the SCC allowed a greater (60%) and longer lasting (at least 1 hour) VsEP amplitude reduction. Thus, this study supports the hypothesis that the vestibular end organs become more sensitive to sound stimuli (Tullio phenomenon1) in the presence of a fenestrated SCC, which has been reported in other studies.14,15 The clinical findings in the patients with dehiscence over the superior semicircular canal9 also demonstrate sensitivity to sound and pressure stimuli: loud sounds in most and changes in middle ear pressure in about half induced vertigo or oscillopsia. The symptoms resolved or improved following surgical procedures to correct the dehiscence in 5 of the 17 patients.9 These clinical and surgical findings provide additional support for the suggestion that fenestration or dehiscence of a SCC allows sound and pressure stimuli to reach and affect the vestibular end organs. In fact, Ribaric et al21,22 have taken clinical advantage of this by fenestrating the SCC in patients with profound hearing loss and normal vestibular function, hoping that such a procedure would enable the patients to make use of the vestibular end organs to perceive sound stimuli. Following fenestration, the patients reported improved hearing to bone-conducted, not air-conducted, stimuli.
In conclusion, fenestration of the SCC makes the vestibular end organs more sensitive to acoustic stimulation, presumably by shifting additional acoustic energy from the cochlea to the vestibule. This finding may have clinical significance.
Accepted for publication October 2, 2001.
Corresponding author: Haim Sohmer, PhD, Department of Physiology, Hebrew University-Hadassah Medical School, PO Box 12272, Jerusalem, Israel (e-mail: firstname.lastname@example.org).