Measuring the cochlear blood flow and distortion-product otoacoustic emissions during reversible cochlear ischemia: a rabbit model

Hearing Research 133 (1999) 40^52

Measuring the cochlear blood £ow and distortion-product otoacoustic emissions during reversible cochlear ischemia: a rabbit model Thierry Mom b

b;

*, Fred F. Telischi a , Glen K. Martin a , Brenda L. Lonsbury-Martin

a

a Department of Otolaryngology, University of Miami School of Medicine, Miami, FL, USA Service d'Otorhinolaryngologie et Laboratoire de Biophysique Sensorielle (2R3), Faculte¨ de Me¨decine, Universite¨ D'Auvergne, BP 38, 63001 Clermont-Ferrand, France

Received 11 August 1998; received in revised form 27 January 1999; accepted 14 March 1999

Abstract Impairment to the cochlear blood flow likely induces many types of sensorineural hearing loss. Models using several small laboratory animals have been described in the literature that permit the simultaneous monitoring of the cochlear blood flow with laser-Doppler flowmetry and cochlear function using evoked responses. However, these models have not permitted a direct application of the resulting knowledge to the human condition, primarily due to differences in the translucence of the otic capsule between species. In the present study, to approximate conditions relevant to the human patient, the rabbit was utilized to develop a procedure in which laser-Doppler flowmetry could be used to measure the cochlear blood flow in an animal with an opaque otic capsule. At the same time, the cochlear function was monitored non-invasively using distortion-product otoacoustic emissions. In this manner, a laser-Doppler probe was positioned in the round window niche and the cochlear function measured using distortionproduct otoacoustic emissions during a systematic series of ischemic episodes. Cochlear ischemia was produced by deliberately compressing the eighth nerve complex at the porus of the internal acoustic meatus, for periods lasting from 1^3 min, while cochlear blood flow and distortion-product otoacoustic emission measures were obtained simultaneously before, during and following the occlusion. Results demonstrated that the cochlear blood flow sharply decreased within 1 s after compression onset, whereas distortion-product otoacoustic emissions showed obstruction-induced changes after a delay of several seconds, provided that the blood flow decreased, at least, 40%. Similarly, upon release of the compression, the cochlear blood flow began to recover within 1 s, whereas the recovery of the corresponding distortion-product otoacoustic emissions was slightly delayed. Although not apparent in the distortion-product otoacoustic emission recovery time course, the cochlear blood flow consistently overshot its initial baseline value during the recovery process. Thus, although cochlear ischemia produced changes in the distortion-product otoacoustic emission activity that generally followed the resulting alterations in the cochlear blood flow, the detailed relationship between the two measures was complex. ß 1999 Elsevier Science B.V. All rights reserved. Key words: Cochlear blood £ow; Laser-Doppler £owmetry; Reversible cochlear ischemia; Distortion product otoacoustic emission; Rabbit

1. Introduction Measuring the cochlear blood £ow (CBF) has been of experimental interest for a number of years, mainly because cochlear ischemia is assumed to be one of the principal causes of sensorineural hearing loss (SNHL). Indeed, it is likely that cochlear ischemia forms the

* Corresponding author. Tel.: +33 (473) 62 5549; Fax: +33 (473) 62 5910; E-mail: [email protected]

basis for certain types of presbycusis as well as for many cases of sudden idiopathic SNHL (Schuknecht, 1974). Additionally, some of the hearing loss encountered in patients with acoustic neuromas may also be caused by cochlear ischemia (Bon¢ls and Uziel, 1988) and, most probably, vascular compromise of the cochlea (Telischi et al., 1996) explains the post-operative SNHL that sometimes occurs following surgery in the region of the cerebellopontine angle (CPA). In the interests of discovering an optimal treatment for cochlear ischemia, a number of animal models have

0378-5955 / 99 / $ ^ see front matter ß 1999 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 5 9 5 5 ( 9 9 ) 0 0 0 5 6 - 8

HEARES 3236 17-6-99

T. Mom et al. / Hearing Research 133 (1999) 40^52

been developed to promote a better understanding of the resulting cochlear pathology and dysfunction. For example, in an early study in guinea pigs of the e¡ects of cochlear ischemia, Perlman et al. (1959) showed that transient vascular occlusions caused damage primarily to the outer hair cells (OHCs) and spiral ganglion cells. Because of the technological limitations of that time, it was not possible to objectively evaluate the e¡ects of these ischemic episodes on CBF. However, Perlman et al. (1959) showed a compression-related decrement in CBF subjectively by microscopically observing a decrease in the £ow of blood through individual vessels of the stria vascularis. Several decades after these original observations, laser-Doppler £owmetry was reported to provide accurate and minimally invasive measures of the blood £ow, in general (Stern et al., 1977), and, shortly after its initial discovery, successful measures of CBF using laser-Doppler system were accomplished (Miller et al., 1983). The laser-Doppler systems that are currently available in the marketplace make it possible to routinely measure CBF in experimental animal models including the guinea pig (Miller et al., 1983; Randolf et al., 1990 ; Levine et al., 1993), rat (Scheibe et al., 1990; Seidman et al., 1991) and gerbil (Ren et al., 1995; Mom et al., 1997). In such small laboratory species as gerbils and guinea pigs, the laser beam of these instruments can be easily directed towards the capillary bed of the stria vascularis due to the translucency of the cochlea's bony capsule. Rats are less frequently used, in part, because their otic capsule is thicker and also more dif¢cult to approach. In other animal species, such as rabbits, in which the cochlea is embedded in dense temporal bone, the laser beam is greatly attenuated, thus adversely compromising its measurement capabilities. Primarily due to this technical limitation, there currently are no reports on laser-Doppler measures of CBF in rabbits during experimental transient ischemia. However, Asami et al. (1995) recently reported the e¡ects on CBF of inhaling high carbon dioxide/low oxygen gas mixtures in the resulting hypercapnic rabbit. These authors, using a custom-built laser-Doppler probe to measure CBF from a promontory position, reported an overall decrease of 15^30% in CBF during hypercapnia. However, in the absence of related explicit changes in CBF, as well as the presence of appreciable noise £oor (NF) levels, the validity of their measures is unknown. It is also noteworthy that an independent measure of the cochlear function was not performed in this study as a control for de¢nitively establishing the e¡ects of hypercapnia on CBF. In humans, to date, only a few experimental ¢ndings on laser-Doppler measures of CBF are available (Scheibe et al., 1990; Miller et al., 1991). For example,

41

the ¢ndings of Scheibe et al. (1990) indicated that CBF was measurable by positioning the probe directly on the promontory of the cochlea. In this position, a transient decrease in CBF was noted during breath holding. Scheibe et al. (1990) also reported that the thick cochlear bone of the human attenuated the laser beam up to four times more than did the more translucent otic capsule of the guinea pig. In addition, it is likely that the mucosal vasculature of the promontory, which was in contact with the probe, contaminated their CBF readings in some cases. In a later study, Miller et al. (1991) also placed a laser-Doppler probe on the promontory of a few patients in order to measure CBF under several conditions. Speci¢cally, they showed that CBF changed more dramatically during irrigation of the external ear canal with either warm or cold water than during electrical stimulation of the cochlea using an electrode on the round window membrane. However, Miller et al. (1991) cautioned that movement-related artefacts, which occurred mostly during the irrigation procedure, could have also contributed to the robust changes in CBF which they observed. Nonetheless, these authors were con¢dent that their careful measurements avoided such contamination of the data. Again, in neither human study was cochlear function monitored simultaneously with CBF measurements, nor, for obvious ethical reasons, was the deliberate induction of transient ischemia performed to test the validity of CBF indices. One opportunity to assess the sensitivity of laserDoppler measures of CBF in human cases of cochlear ischemia would be to monitor the blood £ow to the cochlea during surgery involving the CPA, when, at times, an unplanned interruption in CBF can accidentally occur. However, to our knowledge, there have been no such reports, mostly due to the fact that, until recently, it has been impossible to obtain stable CBF baselines (BLs) throughout such long surgical procedures. Also, for the reasons noted above, the opacity of the cochlear bone over the promontory region likely precluded any sensitive laser-Doppler measurements. Although thinning the bone of the promontory by means of a surgical drill would likely increase the sensitivity of the laser-Doppler system to CBF, this strategy is unacceptable risky for preserving the patient's hearing capacity. The most recent animal studies on the e¡ects of cochlear ischemia on the inner ear function have used distortion-product otoacoustic emissions (DPOAEs) as a sensitive index of cochlear activity. These studies performed in rabbit (Widick et al., 1994) and gerbil (Ren et al., 1995; Mom et al., 1997) demonstrated that DPOAEs were extremely sensitive to episodes of cochlear ischemia. In particular, Mom et al. (1997), using a combination of Doppler-based CBF, cochlear electrical

HEARES 3236 17-6-99

42

T. Mom et al. / Hearing Research 133 (1999) 40^52

potentials and DPOAE measures in the gerbil, showed that ischemic-induced alterations in the emissions more accurately re£ected the initial changes in the cochlear function than did some of the more conventional cochlear potentials (i.e. cochlear microphonic, compound action and summating potentials). Presumably, the sensitivity of DPOAEs to the vascular compromise of the cochlea supports earlier ¢ndings that the OHCs of the organ of Corti are early targets of cochlear ischemia (Perlman et al., 1959; Billet et al., 1989). One interesting observation in the Mom et al. (1997) study was that during the post-ischemic recovery of cochlear function from long duration ischemia, CBF returned rapidly to its corresponding BL value, while the recovery of DPOAEs proceeded non-monotonically in that they exhibited a brief secondary reduction in the level before returning to their pre-occlusion BL levels. The primary goal of the present study was to extend the prior description of DPOAEs in rabbits during transient cochlear ischemia by measuring CBF simultaneously using laser-Doppler £owmetry. In this manner, the relationship of ischemic-induced changes in DPOAEs was directly related to corresponding alterations in CBF by performing systematic measurements of the two response parameters before, during and following deliberate occlusions of the blood supply to the cochlea. By developing a reliable method to measure CBF in the rabbit, which, like the human, has a dense cochlear bone, it was also expected that the technique would be amenable to simple modi¢cations that would permit its use as an intra-operative monitor in patients undergoing surgery involving the CPA. A secondary interest concerned determining whether the rabbit, like the gerbil, exhibited a transitory reversal in the recovery of DPOAEs during the immediate post-ischemic period, at the same time that CBF recovered in a more straightforward monotonic manner. Establishing such similarities between the two laboratory models would ensure the generality of the earlier measures across animal species. 2. Methods The overall investigation consisted of two parts involving a preliminary study, which was performed on ¢ve rabbits, and a ¢nal set of experiments that was conducted in 11 animals. In the latter study, 11 rabbits (eight females, three males) of the New Zealand strain (nine pigmented, two albino), ranging in weight from 2 to 4.5 kg, were used as experimental subjects. The protocol for the care and use of the rabbits was reviewed, approved and monitored by the Institutional Animal Care and Use Committee of the University of Miami School of Medicine.

A surgical level of anesthesia in the presence of spontaneous respiration was induced in all rabbits by an intramuscular (IM) injection of a mixture of ketamine hydrochloride (50 mg/kg) and xylazine hydrochloride (10 mg/kg). The IM injections were repeated as needed according to the elicitation of an eye-blink re£ex, thus ensuring that the experiments were performed under adequate anesthetic conditions. All subsequent procedures were performed within a walk-in, sound-treated chamber. The core-body temperature was maintained around 38³C using a thermostatically controlled heating blanket with feedback provided by a rectal thermometer. The head was secured by a surgically implanted headmount device so that it could be ¢rmly oriented at any given angle. The subsequent surgical steps were performed to satisfy two goals. The ¢rst was to expose the cochlea using a post-auricular approach to the middle ear so that the sensor probe of the laser-Doppler system could be placed in a position that would allow optimal monitoring of CBF, whereas the second aim was to access the porus of the internal acoustic meatus in order to permit systematic occlusions of the blood supply to the inner ear. Toward these ends, under the operating microscope, the middle ear was entered postauricularly through the auditory bulla and the dorsal neck muscles were sectioned and resected at the midline to also provide adequate access to the basal skull region. The posterior wall of the bulla was then completely exposed by severing the attachments of the neck muscles. During the remainder of the experiment, the origin of the jugular vein, which was just medial to the bulla, was protected by a strategically placed piece of aluminum foil. In the initial experiments, an electrical dental drill ¢tted with a diamond burr was used to enter the middle ear cavity just below the horizontal border, between the bulla's thick and thin posterior walls. At this time, the upper portion of the wall was gently resected, until the cochlea was widely exposed, using either the diamond burr or rongeurs and a curette. With this exposure, the hook region of the cochlea was clearly visible through the round window membrane. In the pilot rabbits, the probe of the laser-Doppler system was set ¢rmly onto the basal cochlear bone in various positions that approximated the location of the stria vascularis. However, with these placements, the measured CBF values were so low that the resulting limited dynamic range between the blood £ow and NF readings made it di¤cult to distinguish between normal CBF and that associated with ischemic conditions. Thus, a placement site was developed for the probe using the round window niche. Because of the greatly increased dynamic range between normal and ischemic CBF states as well as evidence that, in this position, sensitive and reliable CBF readings were ob-

HEARES 3236 17-6-99

T. Mom et al. / Hearing Research 133 (1999) 40^52

tainable, the round window niche position was used for the remaining rabbits. Using a heavy duty micromanipulator, the probe was positioned within the niche and aligned with the spiral modiolar artery at its origin by positioning it at an angle that pointed the laser beam slightly upward and in a medial direction that was approximately 45³ from the vertical plane. Immediately preceding the placement of the probe, the edges of the resected bulla were sealed with bone wax to avoid any blood droplets dripping onto its tip. Finally, before proceeding with the surgical procedures necessary to expose the CPA, several pieces of surgical gauze were positioned between the probe and the bulla bone to ensure a satisfactory mechanical stability of the CBF measurement system, as well as a dry middle ear. Once the laser probe was set satisfactorily in place, the CPA was exposed through the sub-occipital craniotomy developed for the rabbit by Widick et al. (1994). Brie£y, after rongeuring a large portion of the posterior calvarium, the cerebellum was gently retracted medially to visualize the porus of the internal acoustic meatus and a specially fabricated micropipet was used to deliberately compress the exposed eighth nerve complex. In this way, the blood supply to the cochlea was e¡ectively compromized. Following development of the general surgical and combined CBF/DPOAE measurement procedures in pilot animals, the formal acquisition of response measures, as noted above, was initiated in the remaining 11 rabbits. Speci¢cally, to demonstrate that the round window niche provided a better probe-sensing position, systematic comparisons were made in three rabbits between this location and the one against cochlear bone. In addition, to ensure that acoustic stimuli alone did not a¡ect the blood £ow, CBF readings were obtained systematically for ¢ve animals during the collection of a DP-gram. For this latter purpose, data points were plotted as a function of the time and their average values compared for three consecutive 2.5-min intervals representing the time before, during and following determination of the DP-gram. Finally, to better establish the relationship between compression-induced changes in DPOAEs and CBF, a number of reversible ischemic episodes lasting about either 1 or 3 min were deliberately performed. In order to fully demonstrate the ability of DPOAEs to track the compression-induced changes in CBF, an additional subset of occlusions was completed using the 3-min protocol. Speci¢cally, four ischemic episodes were administered in separate rabbits by introducing a brief recovery period in the middle of the compression interval through releasing the blockage of the eighth nerve complex, until CBF returned to its BL value, and then re-applying the occlusion once more (see example in Fig. 4). Using this protocol, which was always

43

performed as the ¢nal ischemic episode within the series of occlusions administered to each of these four animals, the release of the initial blockage occurred after V1 min of compression, thus permitting the cochlea to recover over a time interval of about 30 s. After this recovery interval, the obstruction was continued for the remaining 1.5 min of the 3-min compression period. 3. Experimental protocol Prior to the onset of the compression period, BL values of both CBF and DPOAE levels and phases were obtained over a period of one to several minutes as described below. In addition, CBF/DPOAE measures were also acquired simultaneously during each occlusion episode. Finally, to allow the cochlea to recover from an ischemic event, the compression was released and the post-ischemic time course of CBF/DPOAE responses was tracked for about 10 min. Speci¢c data consisted of determining the delay in seconds for CBF and the DPOAE level and phase to change by more than 2 S.D.s from BL values at the onset of the obstruction and to return to within 2 S.D.s of their corresponding pre-occlusion readings during the post-compression recovery period. Other measures that were derived from these data included using the spreadsheet software described below to determine the slopes of both the onset and o¡set functions in order to specify the rates of change in the response measures during these critical ischemic stages. For every rabbit, after a systematic series of transient ischemic episodes were performed, the entire eighth nerve bundle, including the labyrinthine vessels, was completely severed at the level of the porus to ensure that, in the absence of CBF, the laser signal decreased to its NF. Lastly, the rabbit was killed using an overdose of anesthetics and a ¢nal measure of CBF was obtained. 4. Parametric measures 4.1. CBF To measure CBF, a commercial device (Vasamedics, LaserFlo BPM2 ), with a wavelength of 780 nm and an optical power of 2 mW at the probe tip, was used. The instrument's rigid needle sensor probe (P433-1), with a tip size of 0.8U10 mm, directed the laser beam. The probe ¢t adequately into the round window niche of the rabbit, which was only slightly larger than the probe tip. The sample rate for acquiring the CBF data was 2 points/s, with the on-line display consisting of a run-

HEARES 3236 17-6-99

44

T. Mom et al. / Hearing Research 133 (1999) 40^52

lowing the preparatory surgery, i.e. prior to the ¢rst ischemic episode, to ensure that the prior surgery on the ear and skull had not adversely a¡ected cochlear function, and at the end of the experiment, to verify that the multiple ischemic events produced only reversible e¡ects on the cochlear activity. To measure the DP-gram, DPOAEs in response to primary tones of three equilevels (L1 = L2 = 45, 60, 75 dB SPL) were obtained as a function of the frequency by sweeping the geometric means (GMs) of the primary tones, i.e. the frequencies described by the equation (f1 Uf2 ):5 , from 1.5 to 11.3 kHz. At these moderate sound levels, the generation site for this particular DPOAE in the rabbit is thought to be around the GM (Martin et al., 1987). Thus, the DPOAE frequencies referred to here are in reference to the GM frequency. The second stimulus paradigm was used to monitor the e¡ects on the cochlear function of the surgical steps used both to expose the cochlea through the middle ear to guarantee successful placement of the probe and the internal acoustic meatus to adequately uncover the eighth nerve as it coursed towards the brainstem. With this protocol, DPOAE levels were continuously measured as a function of time in response to primary tones that were ¢xed at a GM of 9 kHz, representing a sensitive frequency region in the rabbit, and alternated systematically between the three stimulus levels used for the DP-grams. This particular paradigm was sensitive to the associated decrements in cochlear function, that could be long lasting, particularly when a drill was used to penetrate the bulla. In the third stimulus paradigm, which was used to monitor the DPOAE changes during the three stages of an ischemic episode, i.e. pre-, during and post-occlusion, primary tone levels were set at the lowest level of 45 dB SPL and the levels and phases of the DPOAEs, which were obtained in response to primaries at the GMs of either 8 or 10 kHz, were also plotted as a function of time. Again, these particular monitoring

ning average that was computed over a 10-s interval. All CBF data points were stored for o¡-line processing. It is important to note that the £ow units for CBF were provided by the manufacturer in terms of relative units of mlLD /min/100 g of tissue, with the subscript `LD' de¢ned as the relative measurement volume of the system. Because the Doppler equipment was not calibrated for speci¢cally measuring CBF, the units described below were designated as arbitrary units (AUs). After the eighth nerve was severed, CBF values corresponded to the NF of the preparation. In this particular experimental model based on the rabbit, NF included the blood £ow associated with the round window membrane, but did not include that of the cochlear bone. 4.2. DPOAEs DPOAEs at 2f1 3f2 were elicited by f1 and f2 primary tones that were generated by a 16-bit digital signal processor (DSP) (Digidesign Audiomedia) on-board the controlling personal computer system. The primary tones, at an f2 /f1 ratio of 1.25, were delivered to the ear through sound tubes from separate earphones (Etymotic Research, ER-2). A number of samples (n = 8) of the ear canal sound pressure in the presence of the primaries, which were measured with a microphone assembly (Etymotic Research, ER-10), were synchronously averaged by the DSP with an averaging time of 0.5 s. The acoustic probe, containing the sound tubes and the sub-miniature microphone unit, was securely ¢tted to the bony portion of the external ear canal using a modi¢ed foam ear plug (EAR). To measure DPOAEs, several stimulus paradigms were used. First, DPOAE levels were measured as DP-grams, i.e. the DPOAE level as a function of the primary tone frequency, in order to provide a series of general control measures. Toward this end, DP-grams were determined at three time points during each experiment including pre-surgically, as a BL measure, fol-

Table 1 Distribution of various compression protocols across rabbits Rabbits (R = right ear/L = left ear)

Type of compression 1-min ischemia

1-R 2-L 3-L 4-R 5-R 6-R 7-R 7-L

3-min ischemia

1 2 2 0 1 0 1 0 0 1 1 1 0 1 0 2 n=6 n=7 Total compressions = 17

HEARES 3236 17-6-99

3-min ischemia with mid-course release 0 1 0 0 0 1 1 1 n=4

T. Mom et al. / Hearing Research 133 (1999) 40^52

45

Fig. 1. Changes in DPOAE levels and CBF during each occlusion lasting for V1 (A) and V3 min (B). DPOAE levels exhibited less erratic £uctuations between subjects than did CBF rates, thus showing the capability of the cochlea to tolerate some modi¢cations to CBF. Note the delay required for DPOAE levels to decrease after the onset of the blockage (vertical dashed line at time 0), which likely corresponded to the cochlea's ability to be somewhat resistant to ischemia. After the release of compression, CBF consistently overshot its BL value, while this observation was not re£ected by the more monotonic nature of the recovery in the DPOAE level. In the lower plot of B, it is clear, for one occlusion episode (lowest curve), that CBF did not recover. Note that the corresponding DPOAEs shown in the top plot (lowest curve) also did not recover to pre-occlusion BL levels, thus illustrating the excellent ability of DPOAEs to re£ect de¢ciencies in CBF. C

frequencies were selected because they represent the most sensitive and robust DPOAEs in the rabbit. 4.3. Combined CBF/DPOAE measures Before performing an occlusion experiment, the timing clocks of the laser-Doppler device and the microcomputer's DSP board were synchronized. DPOAE and CBF measures for the three stages of each ischemic episode were then plotted as a function of time using the same time scale to determine the consequences of deliberately obstructing the blood supply to the cochlea. 4.4. Statistical analyses The CBF and DPOAE data were transferred electronically to a commercially available spreadsheet (Microsoft, Excel, v. 5.0). This software was used to compute descriptive statistics specifying the means and þ 1 S.D. of the changes in CBF between the cochlear bone and round window niche sites and of the changes in CBF and DPOAE levels and phases between the various stages of an occlusion episode. Changes in the measured parameters were considered signi¢cant when they exceeded 2 S.D.s of their initial average value. Because the DPOAE phase tended to destabilize during ischemia, it was considered to be unstable when its variation between two consecutive points was more than 4 S.D.s of its initial average value. The slopes of the occlusion onset and o¡set time courses, between time points selected to represent the monotonic portions of these functions, were also determined by spreadsheet software. Parametric analyses using a commercial software package (Abacus Concepts, Statview v. 4.5) established statistical di¡erences between CBF recording sites with a paired t-test and between CBF and DPOAE recovery times and rates, for the 1- and 3-min blockages, and for the double compression protocol using a one-way analysis of variance (ANOVA). The adopted level of statistical signi¢cance was P 6 0.05.

5. Results 5.1. Overview During the performance of the study, three of the 11 rabbits died unexpectedly during an early stage of the experiment, while a fourth animal showed a profound loss of DPOAEs immediately following bulla surgery. Thus, the following data are reported for seven rabbits

HEARES 3236 17-6-99

46

T. Mom et al. / Hearing Research 133 (1999) 40^52

(four female, three males) and eight ears, with the ¢nal subject contributing data from two ears. Table 1 presents details concerning the distribution amongst the eight ears of the short lasting 1-min and longer lasting 3-min compression events, along with information concerning which rabbits received the protocol consisting of the double occlusion and intervening recovery period. Overall, the eight ears received six 1-min, seven 3-min and four double occlusion episodes, resulting in a total of 17 compressions. Descriptive statistics specifying the average initial compression-induced changes in CBF and DPOAEs were based on all 17 blockages, because of the identical treatment they received at obstruction onset, regardless of the duration of the ensuing occlusion period. However, the corresponding summary statistics for describing the subsequent recovery time courses at compression o¡set were computed according to the number of episodes of each occlusion type. Speci¢cally, because one of the 3-min blockages produced irreversible e¡ects on CBF, the average recovery time courses for the 1- and 3-min blockages were each based on the outcomes of six compression episodes, whereas all four double blockage experiments provided data for computing the mean recovery time course for this unique compression protocol. 5.2. Preliminary observations Early in the study, it was discovered that the acoustic and vibratory energy of the surgical drill seriously impacted the cochlear function, especially if cutting rather than diamond burrs were used or when the acoustical energy of the drill resonated through a small opening in the bulla. In general, the average losses in DPOAEs at 3^7 dB were greater for the lower frequency region between 3.5 and 4.5 kHz and in response to lower level primaries than they were for the higher frequencies above 8^9 kHz, which showed, on average, a loss of only a few dB. When bone was removed less traumatically with a curette, DPOAE levels across the tested frequency range were much less a¡ected, with the average losses between 3.5 and 4.5 kHz being only a few dB and with no signi¢cant loss in DPOAE levels noted for the higher frequencies. 5.3. General experimental ¢ndings The principal e¡ects of an ischemic episode on the cochlear measures of the blood £ow and DPOAE level are illustrated in Fig. 1A and B, for short (V1 min) and long (V3 min) reversible blockages, respectively. These plots displaying the occlusion-induced changes in DPOAEs (top) and CBF (bottom) as a function of the time re£ect the ¢ndings of all experimental blockages

Fig. 2. For this V3-min (192 s) ischemic episode, the CBF and DPOAE levels and phase changes in response to 45-dB SPL primaries at a GM of 8 kHz were simultaneously obtained as a function of time. The vertical dashed lines in this and the remaining ¢gures demarcate the onset and o¡set of the ischemic episode. The horizontal dotted lines mark the BLs of each depicted parameter. a: CBF in AU; b: DPOAE level in dB SPL and c: DPOAE phase in degrees measured with respect to the primary tones. Note the sequence of compression-induced changes. That is, CBF decreased initially and was closely followed by an increase in the DPOAE phase, which was ¢nally followed by a decline in DPOAE levels. Immediately following compression o¡set, CBF returned to BL and then signi¢cantly increased above pre-compression £ow rates. The recovery of CBF was next followed by the return of the DPOAE level and phase towards BL, except that a small reversal in their recovery time courses was noted around the time at which the recovery rate for CBF slowed. Also, note that the decrease in CBF from its maximum value did not a¡ect the slower secondary monotonic recovery time course of DPOAE levels.

HEARES 3236 17-6-99

T. Mom et al. / Hearing Research 133 (1999) 40^52

Fig. 3. Example of the simultaneous alterations in DPOAE phase and level during a V3-min (189 s) ischemic episode. Following compression onset, after a brief delay, the level monotonically decreased during ischemia while the phase monotonically increased. As the phase exhibited further increases, it rapidly became random while the DPOAE level reached NF values (not shown). During the subsequent recovery process, the phase and level changed in contrasting manners in that as the level again decreased, the phase increased.

and, thus, provide an overview of the detailed ¢ndings presented below. For example, a ¢rst impression from these data is that, because of their less varying trajectories, DPOAEs, when compared to the more erratic compression-induced £uctuations in CBF, were protected somewhat against the adverse e¡ects of abrupt reductions in the blood £ow. Also evident from the plots of Fig. 1 is that the resulting decrements in DPOAE levels consistently occurred with a greater delay from occlusion onset than did the changes observed for CBF. It is also apparent, particularly in Fig. 1B, that the post-occlusion recovery time courses for both CBF and DPOAEs were non-monotonic. The general ¢ndings depicted above in Fig. 1 are more clearly illustrated in the individual example of Fig. 2 where it is clear that after deliberately compressing the eighth nerve complex, for about 3 min, while CBF decreased essentially immediately (Fig. 2a), the resulting reductions in DPOAE level (Fig. 2b) and increases in DPOAE phase (Fig. 2c) occurred after a short delay. The unexpected ¢nding noted above which is illustrated in Fig. 2a was that, following occlusion o¡set, CBF increased to £ow rates of about 12 AU that were appreciably greater than those measured during the corresponding BL period of about 6.5 AU. Moreover, as shown in Fig. 2b, before reaching BL levels during the immediate post-compression period, DPOAEs consistently exhibited a transitory reversal in their recovery time course that caused their levels to decrease again. Frequently, as illustrated in Fig. 3 at about 240 s post-compression onset, the transitory notch in DPOAE level (solid line) during recovery was associated with an increase in phase (open circles). Moreover, this brief delayed reversal in DPOAE recovery was immediately followed by a return toward pre-

47

occlusion levels at a rate that was notably slower than the original recovery rate. Finally, as demonstrated most clearly in Fig. 3 and in a previous study of the e¡ects of transient cochlear ischemia on DPOAEs (Telischi et al., 1998), the phase response exhibited compression-induced alterations before the level parameter. As illustrated around 0 s time of Figs. 4 and 5, it is also noteworthy that in nine of the 17 occlusion episodes, a slight increase in CBF values occurred when the pipette initially touched the eighth nerve complex in the porus, prior to executing a complete compression. The plot in Fig. 4, showing the e¡ects of introducing a recovery period in the middle of the compression interval, illustrates how accurately, overall, the changes in DPOAEs followed the slightly earlier occurring ones in CBF. Clearly, both CBF and DPOAEs faithfully represented the e¡ects on the cochlear status of compressing and releasing the obstruction of the eighth nerve complex. The further details provided below present an explicit accounting of the above-described general ¢ndings. 5.4. Detailed experimental ¢ndings: CBF recordings To illustrate the clear improvement of CBF readings from the round window niche, a systematic comparison was made in three rabbits of BL values taken with the laser-Doppler probe either seated in the niche or abutted against the cochlear bone of the basal turn. This

Fig. 4. Example of a 3-min ischemic episode showing the ability of DPOAEs (open circles, left ordinate) to accurately track the e¡ects of compromising CBF (solid dark line, right ordinate). In this case, V1 min after compression onset, the blockage was released for about 30 s and, as a result, CBF recovered to essentially its BL £ow rate, before a second compression was initiated. It is clear that, following each compression onset, DPOAEs faithfully followed the blockage-induced decrements in CBF, after a short delay. Note that during the ¢nal post-compression period, CBF displayed the consistent OS of its BL £ow rates, whereas DPOAEs exhibited the short lasting reversal in its recovery toward pre-ischemic levels that was illustrated in Fig. 2b.

HEARES 3236 17-6-99

48

T. Mom et al. / Hearing Research 133 (1999) 40^52

contrast clearly showed that CBF values were appreciably greater at 5.9 þ 0.6 AU, for the niche location, than they were at 1.7 þ 0.7 AU, with the probe situated on cochlear bone. With the probe situated in the round window niche, BL CBF values for all eight ears were quite stable over time ranging from 5.5 þ 0.2 to 8.8 þ 0.6 AU. As illustrated in the series of 3-min blockages of Fig. 1B, following the initiation of an ischemic episode by compressing the eighth nerve complex, CBF values sharply decreased. The delay between the occlusion onset and the beginning of the decay in CBF was in the order of 1 s. In actuality, the consistently sharp decrease that accompanied, within 1 s, a successful occlusion onset served as feedback to optimally adjust the pipette position in the porus. On average (n = 17), the slope of the compression-induced reduction in CBF, i.e. the decrement rate, was 0.25 þ 0.1 AU/s. Moreover, once CBF reached the NF, which was, on average, in 16 þ 6 s, it remained stable within 0.63 þ 0.3 AU. As shown in Fig. 2, after release of the compression, CBF, after a delay of about 1 s, showed a rapid initial recovery, followed by an overshoot (OS) in which the blood £ow increased to a rate that was much greater than its corresponding BL value. Not clearly evident in Fig. 2, however, but more apparent in Fig. 5, which

Fig. 5. Schematic presentation of a 3-min ischemic episode showing the detailed after-e¡ects of compression on CBF. The two-stage post-compression recovery time course consisted of a ¢rst phase (S1) involving the rapid increase in CBF to a value at B that was greater than its BL, that was immediately followed by a second slower recovery stage (S2), in which CBF continued to OS its precompression rate until a maximal value (CBFmax ) was reached. A = point in time delimiting the initial CBF reading after compression o¡set; B = point in time when CBF recovery notably changed to a slower rate, before continuing to monotonically increase and BP = point in time after which CBF began to more slowly recover following a brief reversal in its recovery rate to lower values. The slope S1, characterizing the £ow rate during the ¢rst stage of CBF recovery, was calculated between A and B, whereas slope S2, describing the £ow rate during the second stage of recovery, was computed between either B or BP and CBFmax . NF = noise £oor. Note that the slight increase in BL around the blockage onset at 0 s (grey line) represents occasions when CBF sometimes increased momentarily as the pipette was being positioned in the porus.

represents a schematic representation of these data, is the observation that, in many cases, the initial rapid increase in CBF to supra-BL values was immediately followed by a slight decrease in the blood £ow, before it continued its recovery to even greater post-compression CBF values at a much slower rate. To better describe this complex behavior of CBF following compression release, the recovery process was separated into two distinct stages. As shown in Fig. 5, the ¢rst or rapid stage of recovery included the initial post-occlusion increase in CBF to a supra-BL value, before the recovery rate abruptly slowed. The slope of the time course of recovery stage 1 was referred to as S1 and it was calculated from the point in time, called A, that delimited the initial CBF reading after compression o¡set, and the breakpoint B, indicating the time point at which the recovery of CBF abruptly slowed, before it continued toward the maximum CBF value (i.e. CBFmax ). The second stage of recovery, which was slower than stage 1, was primarily characterized by the continuing increase in CBF to values that were signi¢cantly above BL measures. As noted above, at times, the initial portion of the second recovery stage included a brief decrease in the recovery rate that persisted for only a few seconds. However, regardless of the rate of the stage 2 recovery process, eventually, it resulted in a CBFmax recovery value, prior to CBF slowly returning to BL rates. The rate of recovery for stage 2 was described as S2 and this was calculated from the slope of the monotonic portion of the secondary recovery curve between either breakpoint B or BP, depending on whether the CBF recovery time course reversed brie£y before it continued to increase monotonically to the CBFmax point. Thus, S2 was computed between either B or BP(four cases) and CBFmax . The S1, S2, CBFmax and B and BP measurement parameters, along with the BL and NF ones, are all indicated in Fig. 5. The initial rate of the return of CBF towards BL levels, i.e. the value of S1, was grossly dependent on the duration of the blockage. Thus, on average, S1 was faster at 0.37 þ 0.2 AU/s for the 1-min compressions than for the longer 3-min occlusions at 0.29 þ 0.2 AU/s. A similar e¡ect was found for S2, which was, on average, 0.11 þ 0.1 AU/s after a 1-min and 0.05 þ 0.03 AU/s after a 3-min blockage. For the double blockage protocol with the intervening transitory recovery period, the mid-course recovery slope was 0.22 þ 0.1 AU/s on average, while the ¢nal mean S1 recovery slope at 0.30 þ 0.1 AU/s represented a comprise between the 1-min and 3-min compression recovery rates. In addition, the mean S2 value at 0.04 þ 0.02 AU/s approximated the 3-min compression S2 value. The amount of the post-compression OS, i.e. CBFmax , was calculated by computing the di¡erence between the augmented CBF value and its BL level,

HEARES 3236 17-6-99

T. Mom et al. / Hearing Research 133 (1999) 40^52

i.e. CBFmax 3BL, with BL representing the average £ow during the pre-occlusion period. The OS amount was also related to the CBF dynamic range, i.e. the di¡erence between BL and the NF, i.e. BL-NF, in order to further appreciate its signi¢cance. Finally, the relative amount of the OS (i.e. OSR ), with respect to the corrected BL, was calculated using the equation: OSR = (CBFmax 3BL)/(BL3NF). Similar to the observations described above for S1 and S2, the OSR value varied depending on the length of the compression period. That is, on average, it was smaller at 0.71 þ 0.4, for the approximately 1-min ischemic episodes, and slightly greater at 0.75 þ 0.4, for the 3-min blockages, thus indicating that the longer the compression interval, the greater the OS in CBF. The mean OSR in the four double blockage experiments at 0.70 þ 0.12 approximated the comparable value computed for the 1 minocclusion episodes. Thus, these data showed that the CBF was enhanced, on average, by about 71% of its initial rate, for the short lasting compressions, and 75% for the longer ones. Further, the time for CBF to restabilize to its pre-blockage BL value following compression o¡set was, on average, 111 s (range = 49^ 286 s) or 1.9 min, for the 1-min blockages, and 392 s (range = 132^611 s) or 6.5 min, for the 3-min blockages. Finally, after the eighth nerve complex was cut, the laser-Doppler's NF ranged, on average, from 1.9 þ 0.2 to 3.4 þ 0.2 AU, whereas following death, the ¢nal CBF reading indicated that these values rapidly decreased to 0 AU. 5.5. Detailed experimental ¢ndings: DPOAEs Prior to ischemia, DPOAE levels were stable within 1 þ 0.4 dB and phases within 8³ þ 4³ during the BL period. As shown in Fig. 2b, following ischemia onset, the DPOAE level exhibited a sharp reduction that followed the decrease in CBF (Fig. 2a), with an average delay of about 10 þ 5 s. Speci¢cally, on average, DPOAE levels began to decrease after the relative CBF value (i.e. CBF-NF) decreased 70 þ 30%. That is, no signi¢cant DPOAE level changes occurred if CBF decreased less than 40%. Once DPOAE levels began to decrease, further reductions occurred at a rate of 1.2 þ 0.5 dB/s and when NF levels were reached, after about 32 þ 5 s, DPOAEs were not measurable again until CBF recovered. An ischemic change in the DPOAE phase was also noted, as illustrated in Fig. 3. Speci¢cally, the phase rapidly increased following the onset of the compression episode and these changes occurred with a delay that closely followed that of the blockage-induced decay in CBF, i.e. within 4 þ 2 s following the CBF decrease. The DPOAE phase continued to increase monotonically, eventually resulting in a V45³ shift (46 þ 9³), after which it became unstable and rapidly

49

exhibited random values. The rate of the increase in the DPOAE phase change was, on average, 1.7 þ 0.6³/s. In contrast, after release of the compression, the DPOAE level, which followed the recovery of CBF, recovered more quickly than did the phase component of the emitted response. Additionally, on average, the level parameter began to recover toward its BL levels by 8 þ 4.5 s after CBF began to recover. As illustrated in Fig. 2b, the initial recovery process was consistently characterized by a notch representing a secondary decrement in DPOAE levels that occurred following the abrupt slow down in the recovery of CBF, i.e. at the end of the ¢rst recovery stage (i.e. point B in Fig. 5). This notch in DPOAE level was apparent, even at times when CBF did not exhibit a clear momentary decrease, but only showed a sharp reduction in the rate of recovery. After the secondary decrease in DPOAE levels, they increased again, at a slower rate, provided that CBF remained, at least, at the level of its BL value. The recovery process for the DPOAE level was ischemia duration-dependent in that it was shorter following the 1-min occlusions than it was after the 3-min compressions. Speci¢cally, for the 1-min occlusions, on average, DPOAE levels stabilized at their original preischemic values within 30 s (27 þ 7 s) of the recovery onset time. In contrast, for the 3-min occlusions, it took V3 min (186 þ 22 s) for the amplitude to stabilize within 2 dB of their initial level after the onset of their recovery. The DPOAE phase became non-randomized within an average of 10 þ 3 s after blockage release, after which the phase decreased towards its pre-ischemic value. The recovery process for the DPOAE phase was also ischemia duration-dependent in that it returned to BL values in about 1 min, or less (48 þ 24 s), after o¡set of the 1-min blockages, whereas it took V4 min (274 þ 111 s) to reach pre-occlusion levels after the 3-min blockages. As illustrated in Fig. 3, the DPOAE level and phase recoveries were, in general, very similar, apart from the fact that they changed in opposite directions (i.e. level decrease versus phase increase), in that they both exhibited a secondary alteration after the initial recovery, then monotonically proceeded towards their preischemic BL values. 6. Discussion A principal result of the present study was the demonstration that CBF can be sensitively measured by laser-Doppler in the rabbit model, in vivo, when the laser beam was directed through the round window membrane towards the modiolar vessels. This con¢rmation is technically important for both the ¢elds of acoustic research and otological surgical practice. In-

HEARES 3236 17-6-99

50

T. Mom et al. / Hearing Research 133 (1999) 40^52

deed, while the peripheral auditory function using DPOAEs has been previously investigated in rabbits with cochlear ischemia (Widick et al., 1994; Telischi et al., 1998), CBF has not been studied in this species during reversible cochlear ischemia. Thus, by directly relating the e¡ects of compromising the blood supply to the cochlea to the resulting alterations in cochlear function using DPOAEs, the present ¢ndings help us to con¢rm the reliability of emitted responses as early indicators of cochlear ischemia in this species. Indeed, CBF measurements were so reliable that they sensitively provided feed-back concerning the e¡ectiveness of a deliberately applied compression to block the blood £ow. Thus, a successful compromise of the blood supply to the cochlea was clearly indicated by the rapid and sharp decrease in CBF from its stable BL levels. Another indication of the sensitivity of the CBF measure was the results of performing, in several experiments, the reversible compression that was released brie£y in the middle of an ischemic event. These short lasting mid-compression o¡sets were consistently associated with the immediate transitory recovery of CBF toward BL values. It is widely accepted that the blood supply to the cochlea may originate from a number of intra-cranial vascular sources, at least, in some mammalian species (e.g. Randolf et al., 1990; Ren et al., 1994). Thus, in the present study, CBF was interrupted at the same location reported by Mom et al. (1997) in the gerbil, i.e. at the porus of the internal acoustic meatus, which represents, at this anatomical level, the only vascular passage to the cochlea. This location for focusing the occlusion of the cochlea's vascular supply was considered to be optimal because, from where ever they originated, all arteries supplying the cochlea were necessarily obstructed by applying the compression maneuver at this critical site. In support of this reasoning, in the present study, once CBF reached its NF level during an occlusion episode, it remained stable and did not recover like that observed by Ren et al. (1994) during compression of the anterior inferior cerebellar artery, until the obstruction was released. Moreover, by positioning the micropipet directly at the porus under microscopic visualization, compression of the adjacent brainstem was avoided. Thus, this method of compressing the eighth nerve complex likely prevented neural inputs originating in the brainstem from in£uencing cochlear elements, especially in the anesthetized animals in which such feedback mechanisms were likely compromised. Consequently, the post-occlusion OS in CBF rate noted here can only re£ect the post-compression reaction of labyrinthine artery(ies) to the ischemic episode. This CBF enhancement thus provides indirect evidence of the vasodilatation of the labyrinthine vessel(s) after the blockage was removed.

By simultaneously measuring DPOAEs along with the CBF measurements, the cochlear function was also monitored in the present work. In previous studies performed in several species including the rabbit (Widick et al., 1994 ; Telischi et al., 1998) and gerbil (Ren et al., 1995 ; Mom et al., 1997), DPOAEs were shown to be extremely sensitive to cochlear ischemia. In particular, Mom et al. (1997) showed that DPOAEs better re£ected early functional changes in cochlear activity due to compression onset or o¡set than did the conventional electrical potentials of the cochlea including the cochlear microphonic, compound action and summating potentials. The fact that the changes in DPOAEs so dependably followed the compression-induced alterations in CBF further con¢rms the validity of the CBF measures made here as sensitive indicators of the vascular perfusion of the cochlea. The sensitive combined CBF/DPOAE measures showed that DPOAEs altered secondarily to CBF and it was possible to determine, with a great accuracy, the delay of ischemia-related alterations to DPOAEs, by synchronizing the timing mechanisms of the CBF and DPOAE measuring instruments. The earlier work of Telischi et al. (1998) also reported a delay for rabbit DPOAEs to exhibit signi¢cant decrements. Additionally, like that similarly reported in the Telischi et al. (1998) study, the DPOAE phase also changed sooner following compression in the rabbits investigated here than did the corresponding measure of the emission amplitude. These ¢ndings with respect to DPOAE levels are also consistent with what has been previously reported in gerbils, where 60-dB SPL primaries elicited emissions that began to decay within 10 s following ischemia (Mom et al., 1997). However, while in gerbils a two-stage decay in the DPOAE level was observed in which a plateau occurred following an initial rapid decrease, rabbits exhibited, as they did in prior studies (Widick et al., 1994 ; Telischi et al., 1998), only a monotonic reduction in DPOAE levels. This di¡erence between rabbit and gerbil may be species-dependent as it was shown earlier, for example, that the amount of cochlear energy reserve di¡ers between rodent species (Lotz et al., 1977). Di¡erences in primary tone levels (45 dB SPL in rabbits versus 60 dB SPL in gerbils) may also have contributed to the species-speci¢c ¢ndings. A future study is needed to focus on the issue whether di¡erences in species, or the levels of the primaries, contributed primarily to the species unique effects of the beginning stages of cochlear ischemia on DPOAE levels. During the post-blockage period, the initial recovery in CBF, as measured by its slope (S1), was followed by a recovery of DPOAE levels. Moreover, the marked breakpoint (i.e. B) in the CBF recovery curve around BL levels was always followed by a secondary decrease

HEARES 3236 17-6-99

T. Mom et al. / Hearing Research 133 (1999) 40^52

in the DPOAE level, which was consistently mirrored by an increase in the DPOAE phase, whether or not the CBF subsequently exhibited either a brief decrease or a simple change in the rate of recovery. These observations on DPOAEs are consistent with those reported previously for the gerbil, where CBF recovery from reversible ischemia lasting s 1 min was followed by the same secondary reduction in DPOAE levels (Mom et al., 1997). To our knowledge, this represents the ¢rst report addressing the reliability of laser-Doppler £owmetry for measuring CBF during reversible ischemia in rabbits. Rabbits, like humans, have a very dense otic capsule. The demonstrated ability to simultaneously monitor CBF from the round window niche and cochlear function with DPOAEs, while systematically interrupting and re-establishing the blood supply to the cochlea, suggests that these techniques can be e¡ectively used in humans undergoing CPA-related surgery to identify ischemic episodes at a time when their adverse e¡ects may still be reversible. It is interesting that, to date, there are no reports in humans addressing the detection of ischemic-related changes in CBF. One likely reason for this is that aiming the laser beam at the presumed location of the stria vascularis through the dense bone of the cochlea's promontory confounded the earlier measurements of CBF in human patients. It is noteworthy, for example, that Scheibe et al. (1990) suggested that the vasculature of the mucosal layer covering the promontory probably contributed unfavorably to their laser-based CBF measurements. In fact, the preliminary ¢ndings of the present study con¢rmed that when the laser probe was placed directly on the cochlear bone, smaller CBF values were obtained than when it was positioned in the round window niche. Thus, for certain, the opaque bony wall of the cochlea highly attenuates the laser beam's sensitivity to the blood £ow. Moreover, the laser-Doppler's well-established sensitivity to moving elements (Stern et al., 1977) makes it susceptible to the in£uences of all types of movement. Thus, with the probe on the promontory, unexpected involuntary movements may have been an important source of artefacts in previous attempts to obtain stable CBF measures in humans (Miller et al., 1991). The present study demonstrated that when the laser beam was directed through the only translucent part of the rabbit cochlea, i.e. the round window membrane, unvarying and robust measures of CBF were possible. Although this orientation of the probe did not allow measures of CBF from the stria vascularis, it did permit the laser beam to be directed medially towards the spiral modiolar artery, at its origin, from which it presumably detected the movements of red blood cells.

51

This ability to reliably measure CBF through the round window membrane is directly in contrast to ¢ndings previously reported by Miller et al. (1991), who also placed a laser-Doppler probe in the niche of the round window opening. One possible explanation for their lack of success with this approach could be an imperfect alignment of the laser-Doppler probe they used with respect to the main cochlear vessels. It is conceivable that either the size or orientation of their probe was inadequate for aiming the laser beam accurately enough to include the spiral modiolar artery of either the guinea pig or human. In the present study, the size of the probe allowed it to be aligned perfectly within the niche of the rabbit round window so that the blood £ow within the spiral modiolar artery could be successfully measured. In addition, this position of the Doppler probe also allowed it to be better stabilized against the ridges of the round window niche, thus making it less susceptible to non-blood £ow-related artefactual movements. Finally, the experimental rabbit model developed here, which permits laser-Doppler-based measures of CBF, will also allow other types of cochlear dysfunction that have been studied with a purpose in rabbits, including endolymphatic hydrops (Martin et al., 1983, 1989 ; Lonsbury-Martin et al., 1989) and temporary and permanent noise-induced SNHL (Martin et al., 1987; Lonsbury-Martin et al., 1987 ; Mensh et al., 1993a,b), to be characterized with respect to the functional status of the cochlea's vasculature. Obtaining this basic information on CBF may provide some signi¢cant insights into the pathophysiological mechanisms causing either a common hearing problem, like noise-induced hearing loss, or a less common but more puzzling a¥iction, such as Me¨nie©re's disease. In summary, the present study provides evidence that CBF in the rabbit can be sensitively monitored by means of laser-Doppler measures in which the probe is placed within the round window niche. The development of this preparation permits the use of the rabbit as an experimental model to study the role of ischemia in a number of cochlear disorders. In addition, in clinical research, the technique is promising in that it provides new possibilities for monitoring the consequences of CPA surgery on the auditory function in the human patient. Acknowledgements The authors thank Ozcan Ozdamar Ph.D., Barden B. Stagner and Geo¡rey M. Waxman for their assistance with the technical aspects of the study. This work was supported by Grants from the Colle©ge Franc,ais d'Oto-

HEARES 3236 17-6-99

52

T. Mom et al. / Hearing Research 133 (1999) 40^52

Acknowledgements rhinolaryngologie, the SmithKline Beecham Pharmaceutical Company and the United States Public Health Service (DC00613, DC03114). References Asami, K., Nakashima, T., Morisaki, H., Akanabe, K., Kuno, K., Yanagita, N., 1995. E¡ects of hypercapnia on cochlear and cerebral blood £ow in rabbits. ORL 57, 239^244. Billett, T.E., Thorne, P.R., Gavin, J.B., 1989. The nature and progression of injury in the organ of Corti during ischemia. Hear. Res. 41, 189^197. Bon¢ls, P., Uziel, A., 1988. Evoked otoacoustic emissions in patients with acoustic neuromas. Am. J. Otol. 9, 412^417. Levine, R.A., Brown, M.C., Bu-Saba, N., 1993. Laser-Doppler measurements and electrocochleography during ischemia of the guinea pig cochlea: Implications for hearing preservation in acoustic neuroma surgery. Ann. Otol. Rhinol. Laryngol. 102, 127^136. Lonsbury-Martin, B.L., Martin, G.K., Coats, A.C., Johnson, K.C., 1989. Alterations in acoustic-distortion products and single nerve ¢ber activity in hydropic rabbits. In: Nadol, J.B. Jr. (Ed.), Second International symposium on Me¨nie©re disease: Pathogenesis, Pathophysiology, Diagnosis and Treatment, Kugler and Ghedini, Berkeley, CA, USA, pp. 337^341. Lonsbury-Martin, B.L., Martin, G.K., Probst, R., Coats, A.C., 1987. Acoustic distortion products in rabbit ear canal. I. Basic features and physiological vulnerability. Hear. Res. 28, 173^189. Lotz, P., Haberland, E.J., Kuhl, K.D., 1977. The e¡ect of glycogen in the cochlea metabolism. In: Portmann, M., Aran, J.M. (Eds.), Les Colle©ges de I'Institut National de la Sante¨ et de la Recherche Me¨dicale. Inner ear biology/Biologie de l'oreille interne, Vol. 68, INSERM, Bordeaux, France, pp. 233^242. Martin, G.K., Shaw, D.W.W., Dobie, R.A., Lonsbury-Martin, B.L., 1983. Endolymphatic hydrops in the rabbit: Auditory brainstem responses and cochlear morphology. Hear. Res. 12, 65^87. Martin, G.K., Stagner, B.B., Coats, A.C., Lonsbury-Martin, B.L. (1989) Endolymphatic hydrops in rabbits: Behavioral threshold, acoustic distortion products, and cochlear pathology. In: Nadol, J.B., Jr. (Ed.), Second International symposium on Me¨nie©re disease: Pathogenesis, Pathophysiology, Diagnosis and Treatment, Kugler and Ghedini, Berkeley, CA, USA, pp. 205^219. Martin, G.K., Probst, R., Scheinin, S.A., Coats, A.C., Lonsbury-Martin, B.L., 1987. Acoustic distortion products in rabbits. II. Sites of origin revealed by suppression and pure-tone exposures. Hear. Res. 28, 191^208. Mensh, B.D., Lonsbury-Martin, B.L., Martin, G.K., 1993a. Distortion product emissions in rabbit: II. Prediction of chronic noise e¡ects by brief pure-tone exposures. Hear. Res. 70, 65^72.

Mensh, B.D., Patterson, M.C., Whitehead, M.L., Lonsbury-Martin, B.L., Martin, G.K., 1993b. Distortion product emissions in rabbit: I. Altered susceptibility to repeated pure-tone exposures. Hear. Res. 70, 50^64. Miller, J.M., Marks, N.J., Goodwin, P.C., 1983. Laser Doppler measurements of cochlear blood £ow. Hear. Res. 11, 385^394. Miller, J.M., Bredberg, G., Grenman, R., Suonpa«a«, J., Lindstro«m, B., Didier, A., 1991. Measurement of human cochlear blood £ow. Ann. Otol. Rhinol. Laryngol. 100, 44^53. Mom, T., Avan, P., Romand, R., Gilain, L., 1997. Monitoring of functional changes after transient ischemia in gerbil cochlea. Br. Res. 751, 20^30. Perlman, H.B., Kimura, R., Fernandez, C., 1959. Experiments on temporary obstruction of the internal auditory artery. Laryngoscope 69, 591^613. Randolf, H.B., Jaupt, H., Scheibe, F., 1990. Cochlear blood £ow following temporary occlusion of the cerebellar arteries. Eur. Arch. Otorhinolaryngol. 247, 226^228. Ren, T., Brown, N.J., Zhang, M., Nuttall, A.L., Miller, J.M., 1995. A reversible ischemia model in gerbil cochlea. Hear. Res. 92, 30^37. Ren, T., Avinash, G.B., Nuttall, A.L., Miller, J.M., Laurikainen, E.A., Quirk, W.S., 1994. Dynamic response of cochlear blood £ow to occlusion of anterior inferior cerebellar artery in guinea pigs. J. Appl. Physiol. 76, 212^217. Scheibe, F., Haupt, H., Berndt, H., Magnus, S., Weymar, P., 1990. Laser light transmission and laser Doppler £ow measurements on the human, rat and guinea pig cochlea. Eur. Arch. Otorhinolaryngol. 247, 20^23. Schuknecht, H.F., 1974. Pathology of the Ear. Harvard Univ. Press, Cambridge, MA, USA, pp. 303^317. Seidman, M.D., Quirk, W.S., Nuttall, A.L., Sweitzer, V.G., 1991. The protective e¡ects of allopurinol and superoxide dismutase-polyethylene glycol on ischemic and reperfusion-induced cochlear damage. Otolaryngol. Head Neck Surg. 105, 457^463. Stern, M.D., Lappe, D.L., Bowen, P.D., Chimosky, J.E., Holloway, G.A., Jr., Keiser, H.R., Bowman, R.L., 1977. Continuous measurement of tissue blood £ow by laser-Doppler spectroscopy. Am. J. Physiol. 232, 441^448. Telischi, F.F., Widick, M.P., Lonsbury-Martin, B.L., McCoy, M.J., 1996. Monitoring cochlear function intraoperatively using distortion product otoacoustic emissions. Am. J. Otol. 16, 597^608. Telischi, F.F., Stagner, B.B., Widick, M.P., Balkany, T.J., LonsburyMartin, B.L.,, 1998. Distortion-product otoacoustic emission monitoring of cochlear blood £ow. Laryngoscope 108, 837^842. Widick, M.P., Telischi, F.F., Lonsbury-Martin, B.L., Stagner, B.B., 1994. Early e¡ects of cerebellopontine angle compression on rabbit distortion-product otoacoustic emissions: A model for monitoring cochlear function during acoustic neuroma surgery. Otolaryngol. Head Neck Surg. 111, 407^416.

HEARES 3236 17-6-99