- Email: [email protected]
Delayed low frequency hearing loss caused by cochlear implantation interventions via the round window but not cochleostomy
Hearing Research 333 (2016) 49e57
Contents lists available at ScienceDirect
Hearing Research journal homepage: www.elsevier.com/locate/heares
Research paper
Delayed low frequency hearing loss caused by cochlear implantation interventions via the round window but not cochleostomy David Rowe, Scott Chambers, Amy Hampson, Hayden Eastwood, Luke Campbell, Stephen O'Leary* Otolaryngology, Department of Surgery, The University of Melbourne, 2nd Floor, Peter Howson Wing, Royal Victorian Eye and Ear Hospital, 32 Gisborne St, East Melbourne, VIC 3002, Australia
a r t i c l e i n f o
a b s t r a c t
Article history: Received 20 July 2015 Received in revised form 7 December 2015 Accepted 7 December 2015 Available online 29 December 2015
Cochlear implant recipients show improved speech perception and music appreciation when residual acoustic hearing is combined with the cochlear implant. However, up to one third of patients lose their pre-operative residual hearing weeks to months after implantation, for reasons that are not well understood. This study tested whether this “delayed” hearing loss was influenced by the route of electrode array insertion and/or position of the electrode array within scala tympani in a guinea pig model of cochlear implantation. Five treatment groups were monitored over 12 weeks: (1) round window implant; (2) round window incised with no implant; (3) cochleostomy with medially-oriented implant; (4) cochleostomy with laterally-oriented implant; and (5) cochleostomy with no implant. Hearing was measured at selected time points by the auditory brainstem response. Cochlear condition was assessed histologically, with cochleae three-dimensionally reconstructed to plot electrode paths and estimate tissue response. Electrode array trajectories matched their intended paths. Arrays inserted via the round window were situated nearer to the basilar membrane and organ of Corti over the majority of their intrascalar path compared with arrays inserted via cochleostomy. Round window interventions exhibited delayed, low frequency hearing loss that was not seen after cochleostomy. This hearing loss appeared unrelated to the extent of tissue reaction or injury within scala tympani, although round window insertion was histologically the most traumatic mode of implantation. We speculate that delayed hearing loss was related not to the electrode position as postulated, but rather to the muscle graft used to seal the round window post-intervention, by altering cochlear mechanics via round window fibrosis. © 2015 Elsevier B.V. All rights reserved.
Keywords: Cochlear implant Round window Cochleostomy Hearing loss
1. Introduction Electroacoustic hearing (EAS) has led to improved speech perception and music appreciation in cochlear implant (CI) recipients (Gstoettner et al., 2006; Dorman et al., 2009; Adunka et al., 2013; Gantz et al., 2005; Kiefer et al., 2005; Gfeller et al., 2006). Yet
Abbreviations: EAS, electroacoustic hearing; CI, cochlear implant; ABR, auditory brainstem response; SV, stria vascularis; PTA, pure tone average; LBT, lower basal turn; UBT, upper basal turn; LT2, lower second turn; UT2, upper second turn; IHC, inner hair cell; OHC, outer hair cell; SGC, spiral ganglion cell; OSL, osseous spiral lamina * Corresponding author. E-mail addresses: [email protected] (D. Rowe), [email protected] (S. Chambers), [email protected] (A. Hampson), [email protected]. au (H. Eastwood), [email protected] (L. Campbell), sjoleary@unimelb. edu.au (S. O'Leary). http://dx.doi.org/10.1016/j.heares.2015.12.012 0378-5955/© 2015 Elsevier B.V. All rights reserved.
up to one third of CI recipients experience a progressive loss of residual hearing (Gstoettner et al., 2006; Woodson et al., 2010) which threatens to erode the benefits of EAS. The cause(s) of this hearing loss remain elusive. Delayed hearing loss in humans typically occurs in the months after CI surgery, often following a period of stable hearing. This suggests that delayed hearing loss may be associated with either a chronic or slowly progressing intracochlear pathology. One such candidate is the tissue response to the intracochlear electrode array, which is typified by fibrosis, neoosteogenesis, and an infiltrate of inflammatory cells such as mononuclear leukocytes and histiocytes (Cervera-Paz and Linthicum, 2005; Seyyedi and Nadol, 2014) e findings consistent with this being predominantly a foreign body reaction (Gstoettner et al., 2006; Dorman et al., 2009; Adunka et al., 2013; Gantz et al., 2005; Kiefer et al., 2005; Gfeller et al., 2006; Cervera-Paz and Linthicum, 2005; Nadol et al., 2008; Gstoettner et al., 2004; Ilberg Von
50
D. Rowe et al. / Hearing Research 333 (2016) 49e57
et al., 1999; Turner et al., 2008). Giant cells within a foreign body reaction produce free radicals to degrade the implanted material, and have been implicated in some cases of “soft failure” of CI devices (Gstoettner et al., 2006; Adunka et al., 2013; Seyyedi and Nadol, 2014; Nadol et al., 2008; Choi and Oghalai, 2005). If similar mechanisms occur near the organ of Corti, it is conceivable that the sensory neuroepithelium could be damaged, resulting in delayed hearing loss. Intracochlear fibrosis (or osteoneogenesis) could also potentially influence hearing by interfering with cochlear mechanics (Gstoettner et al., 2006; Woodson et al., 2010; Choi and Oghalai, 2005). Fibrosis within the basal turn of scala tympani is thought to dampen acoustic tuning (i.e. hearing) at the apex of the cochlea, while fibrosis of the basilar membrane is thought to dampen acoustic tuning locally (Cervera-Paz and Linthicum, 2005; Seyyedi and Nadol, 2014; Choi and Oghalai, 2005; O'Leary et al., 2013). Consistent with this, we have observed an association between extensive fibrosis and gradual progression of hearing loss in guinea pigs over the first month after surgery (O'Leary et al., 2013; Shepherd et al., 1983; Xu et al., 1993; James et al., 2008). In light of these considerations, intracochlear fibrosis could conceivably cause delayed hearing loss if its physical characteristics change over time. This study intended to examine whether the trajectory of the electrode array within scala tympani might influence delayed hearing loss. We reasoned that the electrode array trajectory would determine the location and possibly the extent of the foreign body (“tissue”) response, and therefore its effects upon cochlear mechanics or the organ of Corti. The electrode array trajectory was controlled by using the two main surgical approaches to cochlear implantation, namely insertion via the round window (to orient the electrode near to the basilar membrane), or a cochleostomy. To further control electrode array position, we developed a cochleostomy technique that caused the electrode array to traverse scala tympani either more medially or laterally relative to the modiolus. Auditory brainstem responses were recorded at selected time points over a 12 week period, at which time cochleae were harvested then subjected to histopathological analysis.
2. Materials and methods 2.1. Experimental design All experiments were approved by and conducted according to the guidelines of the Animal Research Ethics Committee of the Royal Victorian Eye and Ear Hospital (principal investigator David Rowe; application 11/236AR). The study was performed on 35 Dunkin-Hartley tricolour guinea pigs (350 g). These were randomly allocated to five treatment groups. In two round window treatment groups, the round window membrane was incised, with an electrode array then inserted (round window insertion group) or not inserted (round window incision group) into the scala tympani. The remaining treatment groups all underwent scala tympani cochleostomy, followed in one group by laterally directed insertion of an electrode array (lateral insertion group), or by medially directed insertion of an electrode array in another group (medial insertion group). Auditory brainstem responses (ABRs) were recorded before surgery and 1, 4 and 12 weeks after surgery, after which subjects were euthanised and their cochleae harvested for histological analysis. . All subjects possessed normal hearing (defined here as an ABR threshold to a 100 ms click less than 48 dB peakepeak equivalent) before surgery.
2.2. Auditory brainstem responses Subjects were anaesthetised with a intramuscular injection of mixed ketamine (40 mg/kg) and xylazine (4 mg/kg). The ABR recording system has been described previously (Shepherd et al., 1983; Xu et al., 1993; James et al., 2008; Merchant, 2010). Computer-generated acoustic stimuli (clicks and tone pips lasting 5 ms with 1-ms rise/fall times, with tone pips presented at frequencies of 2, 8, 16, 24 and 32 kHz) were delivered to anaesthetised subjects via a loudspeaker (Richard Allen DT-20, UK) placed 10 cm from the pinna. ABRs were recorded differentially using subcutaneous needle electrodes placed at the vertex and nape. A grounding electrode was placed further caudally in lateral subcutaneous tissue. Ear mould compound (Otoform, Dreve, Germany) was used to attenuate hearing in the contralateral ear. Using this method, the frequency-specific cross-head attenuation (measured from the occluded ear canal of the guinea pig) is 28, 27, 35, 45 and 28 dB at 2, 4, 8, 16 and 32 kHz, respectively. Responses were amplified by a factor of 100,000 (DAM-5A, WeP Instruments Inc., USA) and band-pass filtered (Krohn-Hite 3750, Avon, USA) between 150 Hz and 3 kHz (6 dB/octave). The filter output was fed to a 16- bit analogue-to-digital converter (Tucker Davis Technologies, USA) and sampled at 20 kHz for a period of 10 ms following stimulus onset. Thirty stimuli were presented per second. Responses were averaged over 250 stimulus repetitions. Stimulus intensity was decremented in 5 dB steps from high stimulus levels to sub threshold. Waveforms were exported to a software analysis program (written by Dr. James Fallon, adapted by Prof. Stephen O'Leary using Igor 5.02, Wavemetrics Inc., USA) and analysed by the researcher who was blinded to subjects' identity and treatment group. Threshold was defined as the lowest intensity stimulus that evoked a response of >0.4 mV amplitude in wave III of the ABR (responses of this amplitude reliably exceeded background noise). 2.3. Electrode array The electrode array comprised three platinum rings, each welded to a platinum wire 25 mm in diameter, all housed within a Silastic® carrier (MDX4-4210, Dow Corning Products, USA). The electrode tip was tapered to a diameter of 0.41 mm and had a maximum diameter of 0.45 mm. The three platinum rings, each separated by 0.75 mm, were situated near the tip of the electrode and served as visible markers of insertion depth. The total length of the array was 15 mm. All electrode arrays were sterilised before use. 2.4. Cochlear interventions Guinea pigs were anaesthetised with an intramuscular injection of mixed ketamine (60 mg/kg) and xylazine (4 mg/kg). Local anaesthesia was induced at the target ear with a subcutaneous injection of 2% lignocaine. A curvilinear post-auricular incision was made through skin and subcutaneous tissues to expose the bulla. A #11 blade on a #3 handle rotated in the hand was used to make a bullostomy, through which the round window and basal turn were sighted. Care was taken not to disturb the ossicles or external auditory canal. The cochlea was accessed either by incising the round window membrane with a micro pick (round window groups) or by using a 1 mm diamond bur to drill a cochleostomy in the basal turn adjacent to the round window (cochleostomy groups). In the round window, medial and lateral insertion groups, an electrode array was inserted into the scala tympani until resistance was encountered. The round window or cochleostomy was then sealed with harvested muscle. In the lateral and medial
D. Rowe et al. / Hearing Research 333 (2016) 49e57
insertion groups, the muscle was positioned either inferiorly (medial insertion) or superiorly (lateral insertion) to the electrode array, in order to guide the array towards its desired trajectory in scala tympani. In all insertion groups, the distal end of the electrode array was severed with scissors where it exited the bulla then discarded. The dissected subcutaneous tissues and skin were closed with sutures. Subjects were then administered with a subcutaneous injection of analgesic (Temgesic™, Reckitt Benckiser, Australia) and allowed to recover.
51
according to Wada's adaptation (Choi and Oghalai, 2005; Wada et al., 1998) of Greenwood's equations (Greenwood, 1990). 2.7. Statistics All statistical analyses were performed with IBM SPSS Statistics Version 22 (IBM Corp., USA). Data were variously analysed with Pearson's correlation, chi-square test, Mann Whitney U test and ANOVA, the latter incorporating post-hoc tests and applying a Greenhouse-Geisser correction where required.
2.5. Euthanasia and specimen preparation 3. Results At the conclusion of the 12 week ABR, subjects were euthanised under ketamine/xylazine anaesthesia with an intraperitoneal injection of phenobarbitol. After respiratory arrest, subjects were intracardially perfused with 0.9% saline followed by 10% formalin. Cochleae were harvested and immersed in 10% formalin for at least 48 h. Cochleae were then decalcified over 2e3 weeks in 4% ethylenediaminetetraacetic acid before being embedded in Spurr's resin. Specimens were oriented so that the round window membrane was sectioned in a transverse plane. Sections were collected at 125 mm intervals and stained with hematoxylin and eosin. 2.6. Histological analyses Histological analyses were adapted for the guinea pig from a method described by Merchant (2010). Analyses began at the centre of the modiolus (mid-modiolar), at the section that contained the largest extent of cochlear nerve. To define the scope of analyses, we measured the area of the spiral limbus at the middle of the modiolus within each cochlear turn. We then restricted all other histological analyses to cochlear turns in which the area of the spiral limbus was 25% of that recorded at mid-modiolar. This was done because histological structures become distorted in sections approaching the periphery of the cochlea. Analyses were limited to the basal and second turns of cochleae because, according to the Greenwood equation (Greenwood, 1990), these turns encompass areas responsible for sensing 2, 8, 16, 24 and 32 kHz sounds (matching the frequencies of our pure-tone ABRs). Slides were digitally scanned using a Mirax Scan™ (Zeiss, Germany) slide scanner with a 20 optical objective lens. Pannoramic Viewer™ 1.15.2 (3DHistech, Hungary) was used to count inner and outer hair cells and pillar cells, and to estimate the density of cells in Rosenthal's canal, the stria vascularis (SV), and the T1, T2, T3 and T4 regions of the lateral cochlear wall as described by Mizutari et al. (2011). We also scored the condition of the osseous spiral lamina (OSL; 0 ¼ intact, 1 ¼ inner lamina fractured, 2 ¼ both laminae fractured) and the extent to which the basilar membrane was contacted by tissue reaction (i.e. fibrosis and osteoneogenesis) (0 ¼ basilar membrane not in contact with tissue reaction, 1 ¼ minority of basilar membrane in contact with tissue reaction, 2 ¼ most but not all of basilar membrane in contact with tissue reaction, 3 ¼ all of basilar membrane in contact with tissue reaction). All analyses were performed on two or more sections per cochlear turn. Amira™ 5.4.3 (Visualization Sciences Group, France) was used to reconstruct cochleae in three dimensions (Fig. 1a) in order to calculate the volume of tissue reaction and the electrode array trajectory within the basal turn of implanted cochleae. To determine electrode array trajectory, sections were aligned and splines assigned to the centre of electrode tract (which was visible within tissue reaction), the medial and lateral cochlear walls, the basilar membrane, and the organ of Corti. The resulting co-ordinate data were used to calculate the electrode array trajectory relative to the medial and lateral cochlear walls and the organ of Corti. Additionally, Amira was used to map frequencies onto cochleae
3.1. Electrode array trajectory Electrode array trajectories were plotted relative to the medial and lateral cochlear walls, the basilar membrane and organ of Corti (Fig. 1b, c). The horizontal position of the electrode array was expressed as a proportion of the width of the scala tympani, with medial placement as 0 and lateral placement as 1. Electrode array trajectories matched their intended paths, i.e. arrays oriented either medially or laterally maintained these positions within scala tympani. Round window insertions were the closest to the organ of Corti and basilar membrane over the majority of their intrascalar course. The electrode array tip generally ended near the basilar membrane and organ of Corti in 8e10 kHz region of the cochlea, irrespective of whether arrays were located medially or laterally within scala tympani. 3.2. Auditory brainstem responses Post-surgical threshold shifts for all treatment groups are shown in Fig. 2, with pre-surgery thresholds listed in Table 1. In the round window insertion group, thresholds shifts at 8, 16, 24 and 32 kHz were not significantly different over time. In contrast, threshold shifts at 2 kHz increased significantly over time (F1.084, 6.505 ¼ 13.524, p ¼ 0.008, Greenhouse-Geisser correction ε ¼ 0.542, partial h2 ¼ 0.693), averaging 6.3 ± 3.7 dB at 1 week and 22.9 ± 3.4 dB at 12 weeks. Post-hoc comparisons showed that 2 kHz threshold shifts differed significantly between 1 and 12 weeks (p < 0.001), but not between 1 and 4 weeks after surgery (p ¼ 0.752). The pure tone average (PTA) threshold shift (the average threshold shift across all tested frequencies) for the contralateral ear was 4.1 ± 3.5 dB at 12 weeks. Threshold changes in the round window incision group showed a similar pattern over time to that observed for round window insertions. Threshold shifts at 8, 16, 24 and 32 kHz were not significantly different over time. In contrast, 2 kHz thresholds increased from 1 week (11.1 ± 3.1 dB) to 12 weeks (26.1 ± 3.3 dB) post-operatively. This change was statistically significant (F2, 2 16 ¼ 6.595, p ¼ 0.008, partial h ¼ 0.452). Changes at other frequencies were not significant. The PTA threshold shift for the contralateral ear was 7.1 ± 5.6 dB at 12 weeks. Threshold shifts in the medial insertion group either declined or rose then declined over time across all frequencies measured, although these trends were not statistically significant. Therefore, there was no indication of delayed rise in threshold shifts in this group. The PTA threshold shift for the contralateral ear in the medial insertion group was 1.8 ± 1.5 dB at 12 weeks. Threshold shifts in the lateral insertion group tended to decline over time. The PTA threshold shift for the contralateral ear in the lateral insertion group was 1.4 ± 2.6 dB at 12 weeks. Threshold shifts in the cochleostomy only group did not differ significantly over time. The PTA threshold shift for the contralateral ear in the cochleostomy only group was 1.1 ± 5.4 dB at 12 weeks.
52
D. Rowe et al. / Hearing Research 333 (2016) 49e57
Fig. 1. (a) A cochlea from the round window insertion group three-dimensionally reconstructed from digitally scanned sections, showing the electrode array trajectory (blue) surrounded by tissue reaction (yellow/orange) in the basal turn of scala tympani. Such reconstructions were used to calculate the distance of electrode arrays from (b) the medial and lateral cochlear walls, expressed as a ratio from 0 (most medial) to 1 (most lateral), and (c) the organ of Corti, expressed in mm, in the round window insertion (blue), medial insertion (red) and lateral insertion (green) groups. In (b) and (c), bold lines indicate mean distances while dotted lines bound the standard error of the mean. Array trajectory was measured from the centre of the electrode tract, which was visible within scala tympani tissue reaction (fibrosis and osteoneogenesis). The electrode array had a maximum diameter of 0.45 mm.
An ANOVA showed that threshold shifts at 12 weeks did not differ between interventions at 8, 16, 24 or 32 kHz. However, threshold shifts at 12 weeks for the 2 kHz stimulus were significantly different between treatment groups (F4, 33 ¼ 4.457, p ¼ 0.005). A post-hoc analysis revealed a significant difference between threshold shifts in the round window incision group (26.1 ± 9.9 dB) and the lateral insertion group (7.9 ± 8.6 dB), with a difference of 18.2 dB (95% CI, 2.8e33.7, p ¼ 0.014). 3.3. Histopathology and electrode array trajectory The volume of scala tympani tissue reaction for all treatment groups is shown in Fig. 3, with fibrosis and osteoneogenesis presented separately. The three electrode array insertion groups exhibited significantly greater volumes of tissue response (U ¼ 214, z ¼ 2.810, p ¼ 0.004) than the controls (round window incision and cochleostomy-alone), although 44% of cochleae subjected to cochleostomy-alone contained tissue exceeding 10% of the basal scala tympani volume (this was not seen in any round window incision cochleae). The route of electrode array insertion did not have an effect on the volumes of fibrosis (c (Dorman et al., 2009) (2) ¼ 0.097, p ¼ 0.953) or osteoneogenesis (c2(2) ¼ 0.859, p ¼ 0.651). The volume of new bone was correlated with the total volume of tissue reaction within the cochlea (r ¼ 0.659, p ¼ 0.002). Electrode insertion groups showed a similar degree of basilar membrane fixation in the basal turn (c2(2) ¼ 0.189, p ¼ 0.910). Detailed histological analyses were made at the mid-modiolar plane of the lower basal turn (LBT), upper basal turn (UBT), lower second turn (LT2) and upper second turn (UT2) of each cochlea. Repeated measures ANOVAs were undertaken to determine whether histopathology differed with cochlear location (within subjects) or the type of surgical intervention (between subjects). In the organ of Corti, inner hair cell (IHC) counts differed significantly across cochlear locations (F3, 81 ¼ 3.013, p ¼ 0.035). IHC counts were reduced in the basal turn in all three of the cochleostomy intervention groups, but neither of the round window intervention groups. Between groups, surgical intervention approached significance (p ¼ 0.066), and post-hoc testing indicated that IHC counts following round window insertion were significantly lower than those following cochleostomy only (MD ¼ 0.198, p ¼ 0.025), lateral insertion (MD ¼ 0.238, p ¼ 0.015) and medial insertion (MD ¼ 0.261, p ¼ 0.011) (Fig. 4). Outer hair cell (OHC) counts differed significantly across cochlear locations (F3, 81 ¼ 6.820, p < 0.001), with counts lowest in the lower-basal cochlear turn after
all interventions. Surgical intervention had a significant effect on OHC counts (F4, 27 ¼ 2.892, p ¼ 0.041), with post-hoc testing showing that OHC counts were lower after round window insertion than after round window incision (MD ¼ 1.066, p ¼ 0.007), cochleostomy only (MD ¼ 0.916, p ¼ 0.01), lateral insertion (MD ¼ 0.952, p ¼ 0.015) and medial insertion (MD ¼ 0.938, p ¼ 0.021). Pillar cell counts did not differ across cochlear locations (F1.895,53.006 ¼ 1.382, p ¼ 0.260, ε ¼ 0.632), but were dependent upon surgical intervention (F4, 28 ¼ 9.049, p < 0.001), with post-hoc testing revealing that pillar cell counts were significantly lower after round window insertion than after round window incision (MD ¼ 0.736, p < 0.005), cochleostomy (MD ¼ 0.796, p < 0.005), lateral insertion (MD ¼ 0.713, p < 0.005) and medial insertion (MD ¼ 0.896, p < 0.005). Spiral ganglion cell (SGC) densities were lowest in the lowerbasal cochlear turn, and differed significantly across cochlear locations (F2.306, 64.556 ¼ 12.923, p < 0.001, ε ¼ 0.769). On post-hoc testing, SGC densities were higher following round window incision compared with round window insertion (MD ¼ 106.541, p ¼ 0.039) or cochleostomy only (MD ¼ 126.565, p ¼ 0.008), but did not differ significantly from medial or lateral insertions. In the lateral cochlear wall, T4 cell densities differed significantly with surgical intervention (F4, 28 ¼ 7.412, p < 0.001), and were higher for round window insertion than other approaches [round window incision (MD ¼ 712.485, p < 0.001), medial insertion (MD ¼ 604.082, p ¼ 0.001), lateral insertion (MD ¼ 442.393, p ¼ 0.006) or cochleostomy only (MD ¼ 577.649, p < 0.001)] (Fig. 5). Surgical intervention did not affect T1 cell densities (F4, 27 ¼ 2.348, p ¼ 0.080), T2 cell densities (F4, 28 ¼ 1.273, p ¼ 0.304), T3 cell densities (F4, 28 ¼ 1.411, p ¼ 0.256) or SV cell densities (F4, 28 ¼ 0.918, p ¼ 0.467). 3.4. Site-specific histopathology To explore local relationships between histopathological features, these were correlated within each surveyed cochlear region (the lower-basal turn [LBT], upper basal turn [UBT], lower second turn [LT2] and upper second turn [UT2]). Within the organ of Corti, IHC and OHC counts were positively correlated within most cochlear turns (LBT: r ¼ 0.999, p < 0.001; UBT: r ¼ 0.522, p ¼ 0.002; UT2: r ¼ 0.485, p ¼ 0.004). Similarly, pillar cell counts were positively correlated with both IHC counts (UBT: r ¼ 0.576, p < 0.001; UT2: r ¼ 0.642, p < 0.001) and OHC counts (UBT: r ¼ 0.662, p < 0.001; LT2: r ¼ 0.397, p ¼ 0.022; UT2: r ¼ 0.411, p ¼ 0.018). SGN
D. Rowe et al. / Hearing Research 333 (2016) 49e57
53
densities were positively correlated with either IHC counts (UBT: r ¼ 0.473, p ¼ 0.005) or pillar cell counts (LBT: r ¼ 0.599, p < 0.001; UBT: r ¼ 0.425, p ¼ 0.014) in some turns. In the lateral wall, cell densities within one T-cell region tended to correlate (positively) with densities in other regions (Table 2). This applied particularly for T1, T2, T3 and SV regions. Some significant correlations were observed between the histology of the lateral cochlear wall and that of the organ of Corti. T4 cell densities were negatively correlated with OHC counts in the UBT (r ¼ 0.514, p ¼ 0.002) and with OHC and IHC counts in UT2 (OHC: r ¼ 0.370, p ¼ 0.034; IHC: r ¼ 0.450, p ¼ 0.009). T4 cell densities were also negatively correlated with pillar cell counts in each surveyed turn (LBT: r ¼ 0.475, p ¼ 0.005; UBT: r ¼ 0.532, p ¼ 0.001; LT2: r ¼ 0.529, p ¼ 0.002; UT2: r ¼ 0.556, p ¼ 0.001). T2 cell densities were negatively correlated with SGC densities in the LT2 (r ¼ 0.427, p ¼ 0.013), and SV cell densities were negatively correlated with OHC counts in the UBT (r ¼ 0.351, p ¼ 0.045). OHC counts in the LBT were positively correlated with the magnitude of ABR threshold shifts when the latter were averaged across 8, 16, 24 and 32 kHz (r ¼ 0.545, p ¼ 0.001). OSL fracture in the lower basal turn was positively correlated with the density of multinucleated giant cells (MNGC) within each cochlear turn (LBT: r ¼ 0.599, p ¼ 0.001; UBT: r ¼ 0.379, p ¼ 0.030; LT2: r ¼ 0.379, p ¼ 0.030; UT2: r ¼ 0.390, p ¼ 0.044). 4. Discussion 4.1. Overview of results Medially and laterally-directed electrode array insertions followed their expected trajectories. Round window insertion led to an electrode array placement that was closer to the basilar membrane and more centrally located within the scala tympani relative to the medial and lateral cochlear walls. The electrode array tip ended in a similar place regardless of trajectory, most likely because the scalar diameter decreased, constraining the electrode array path. All interventions, including the control groups, produced similar hearing losses when these were averaged across the cochlea, but the round window interventions were associated with a gradual loss of hearing at 2 kHz. Round window insertion caused the most cochlear trauma, being associated with lower OHC and pillar cell counts within the organ of Corti. Round window incision on the other hand caused the least intracochlear trauma. However, both caused a similar degree of delayed, low frequency hearing loss. 4.2. Round window interventions and hearing
Fig. 2. Auditory brainstem response threshold shifts from before surgery to 1 week, 4 weeks and 12 weeks after surgery (mean ± standard error) at each tested frequency. At 12 weeks, the only significant difference in threshold shifts between treatment groups was at 2 kHz. At this frequency, threshold shifts for the round window incision group were significantly larger than those for the lateral insertion group.
Low frequency (2 kHz) hearing deteriorated over time in both of the round window intervention groups, which suggests that this delayed hearing loss was not associated with the presence of the electrode array within scala tympani. Similarly, delayed loss was not a function of the extent of the intracochlear tissue reaction, because there was minimal fibrosis in scala tympani after round window incision, but more extensive fibrosis after electrode insertion (Fig. 3). The extent of injury to the inner ear was lower following round window incision, as evidenced by the greater survival of sensory cells in the organ of Corti (Fig. 4) and lesser effect upon the lateral cochlear wall following this intervention. This suggests that delayed hearing loss may not be related directly to inner ear injury. Rather, delayed hearing loss appears to be associated with the round window intervention itself. This could potentially be related to incision of the round window membrane, or alternatively to the material (muscle) used to seal the round window after surgery.
54
D. Rowe et al. / Hearing Research 333 (2016) 49e57
Table 1 Pre-surgery auditory brainstem response thresholds (mean ± standard error) for each treatment group. Frequency
Threshold (dB SPL) Round window insertion (n ¼ 7)
2 kHz 8 kHz 16 kHz 24 kHz 32 kHz
51.6 39.1 39.0 43.3 62.7
± ± ± ± ±
2.1 4.7 3.6 2.3 3.5
Round window incision (n ¼ 9) 53.6 43.7 40.7 50.1 65.3
± ± ± ± ±
2.6 3.7 3.2 3.1 4.1
Medial insertion (n ¼ 6) 63.8 35.3 29.0 42.3 62.8
± ± ± ± ±
2.7 3.6 2.9 3.6 3.5
Lateral insertion (n ¼ 7) 57.3 32.0 28.3 46.1 60.6
± ± ± ± ±
2.3 2.7 0.7 1.8 2.6
Cochleostomy only (n ¼ 9) 59.1 41.4 39.0 50.7 65.3
± ± ± ± ±
2.6 3.8 2.0 2.8 4.7
Fig. 3. The proportion of the basal scala tympani occupied by (A) fibrosis and (B) osteoneo-genesis (mean ± standard error). The round window, medial and lateral insertion groups (the three groups in which an electrode array was implanted into the cochlea) had significantly more fibrosis than the round window incision and cochleostomy only groups.
Low frequency hearing loss could potentially occur because cochlear mechanics are altered secondary to physical changes at the round window associated with scar maturation after placement of the muscle plug. Choi predicted that scala tympani damping e which would result from reduced round window membrane compliance e would reduce basilar membrane vibrations in the apex of the cochlea (Choi and Oghalai, 2005; Richards, 1981), consistent with the low frequency hearing loss seen here. Further examples of hearing being affected by round window obstruction can be found in the literature. For example, conductive hearing loss has been found to be secondary to congenital absence of the round window membrane in patients. In these cases, removing the excess bone from the predicted location of the round window membrane revealed a functional membrane with some recovery of hearing (Richards, 1981; Eshraghi et al., 2005). While round window membrane damping may have caused the hearing loss, the reason(s) why this was delayed are less clear. We speculate that this may have been due to a change in the compliance of the scar tissue as it matures over time. Another possible mechanism for hearing loss associated with round window surgery is injury or scarring within the middle ear. This is a possibility because of the close proximity between the
Fig. 4. Counts of inner hair cells (IHCs), outer hair cells (OHCs) and pillar cells (PCs) per mid-modiolar histological section in the two most basal cochlear turns (mean ± standard error). Outer hair cell and pillar cell counts differed significantly among treatment groups, with the round window insertion group exhibiting fewer outer hair cells and pillar cells than all other treatment groups.
ossicles and the round window in the guinea pig. Whilst direct contact between the electrode and the ossicles, or scarring of the ossicular chain was not noted during dissection of the cochleae, the histology did not include the ossicles so it is not possible to entirely exclude this possibility. Previous animal studies have used muscle to seal the round window after surgical intervention, but these did not monitor hearing for sufficient time to detect the delayed hearing loss observed here. A study in the rat introduced cochlear electrode arrays via the round window, withdrew them and then sealed the round window with a muscle plug (Eshraghi et al., 2005; James
D. Rowe et al. / Hearing Research 333 (2016) 49e57
55
et al., 2006). This experiment monitored hearing for 7 days after surgery, compared with 12 weeks in our study. A recent guinea pig study inserted dummy CI electrode arrays via either the round window or a cochleostomy, and employed muscle as a tissue seal, in a manner similar to that described here (Burghard et al., 2014). ABR threshold shifts in that study were of a similar magnitude to ours; the investigators did not observe delayed hearing loss, but their observation window was only 28 days, which again was too short to see the delayed hearing loss observed here. It was noted that the ABR thresholds in the contralateral ear of the round window groups were higher than in the cochleostomy groups by a modest degree (between 3 and 4 dB) at the end of the 12 week observation period. Whether this is of real significance is uncertain, given that this difference was not of statistical significance, but it raises the possibility that there may have been a systemic influence on the hearing induced by the round window surgery. 4.3. Cochleostomy interventions and hearing The most notable characteristic of the cochleostomy interventions was that all groups, including cochleostomy without electrode array insertion, exhibited similar ABR threshold shifts. This was surprising, given that previous studies have revealed minimal hearing loss from a small cochleostomy alone, with greater hearing loss arising following electrode array insertion (Braun et al., 2011). We speculate that our results may have been due to the relatively large cochleostomy performed in our study, which was necessary to manipulate the electrode array towards a medial or lateral trajectory. Larger cochleostomies have been associated with greater hearing loss in a clinical setting (James et al., 2008, 2006; Souter et al., 2012; Lee et al., 2013), with larger cochleostomies considered more likely to involve spiral ligament injury, bone dust entering the cochlea, and perilymphatic fluid leakage. It is also possible that in the present experiment, the more extensive muscle plug required to seal the large cochleostomy may have elicited a greater inflammatory response and affected hearing through this mechanism. Despite our concerns that the cochleostomy may have been the major determinant of the hearing loss, several observations are pertinent to the question of delayed hearing loss. In the first month after surgery, the extent of the hearing loss in the cochleostomy groups was very similar to that for the round window interventions (Fig. 2), and most importantly the threshold shifts at 2 kHz did not differ between groups. This means that the cochleostomy models had the sensitivity to detect delayed hearing loss at 2 kHz, so it is still reasonable to conclude that delayed loss is a characteristic of round window, but not cochleostomy interventions. Secondly, the pathology to the organ of Corti associated with the cochleostomy interventions was in general not as extensive as that for round window insertion, and similar or more extensive than round window incision, supporting the interpretation that intracochlear pathology was not the determinant of delayed loss (James et al., 2008; Souter et al., 2012; Lee et al., 2013; XU et al., 1997). 4.4. Cochlear histopathology and electrode array insertion or trajectory Fig. 5. Cell densities in the stria vascularis (SV) and T1, T2, T3 and T4 regions of the lateral cochlear wall (mean ± standard error). Cell densities in the T2 and T4 regions differed significantly among treatment groups. For the T2 region, the round window insertion group had significantly lower cell densities than all other treatment groups. For the T4 region, the round window insertion group had significantly higher cell densities than the cochleostomy, round window incision and medial insertion groups.
The electrode array insertion groups all had significantly more fibrosis in the basal turn of scala tympani than controls, confirming that the presence of a cochlear electrode array promotes a more extensive tissue response. Round window insertion was associated with lower OHC counts and SGN densities across all cochlea regions than either round window incision or the cochleostomy interventions. Following
56
D. Rowe et al. / Hearing Research 333 (2016) 49e57
Table 2 Pearson's correlations (r values) for cell densities in the lateral cochlear wall, across the four most basal cochlear turns. T1 ¼ type 1 fibrocytes, T2 ¼ type 2 fibrocytes, T3 ¼ type 3 fibrocytes, T4 ¼ type 4 fibrocytes, SV ¼ stria vascularis. LBT ¼ lower basal turn, UBT ¼ upper basal turn, LT2 ¼ lower second turn, UT2 ¼ upper second turn. * Correlation significant at 0.05 level (two-tailed). ** Correlation significant at 0.01 level (two-tailed).
T1 cell density
T2 cell density
T3 cell density
T4 cell density
Cochlear turn
T1 cell density
T2 cell density
T3 cell density
T4 cell density
SV cell density
LBT UBT LT2 UT2 LBT UBT LT2 UT2 LBT UBT LT2 UT2 LBT UBT LT2 UT2
1 1 1 1
0.650** 0.597** 0.716** 0.830** 1 1 1 1
0.517** 0.482** 0.484** 0.694** 0.443** 0.389* 0.574** 0.679** 1 1 1 1
0.218 0.603** 0.296 0.237 0.358* 0.495** 0.197 0.250 0.134 0.369* 0.063 0.139 1 1 1 1
0.189 0.407* 0.187 0.529** 0.064 0.407* 0.238 0.569** 0.206 0.513** 0.546** 0.684** 0.281 0.278 0.066 0.003
round window insertion, electrode array trajectories were closer to the organ of Corti, a factor that has previously been associated with greater OHC loss (Choi and Oghalai, 2005; XU et al., 1997). Injury to the organ of Corti does not seem to be due simply to tissue response encroaching on the basilar membrane (Choi and Oghalai, 2005; O'Leary et al., 2013; Lee et al., 2013), since the extent of sensory cell loss was similar across all electrode insertion groups. In the round window insertion group, OHC counts and SGC densities were reduced in the two most basal cochlear turns. In all other groups, OHC counts were only reduced in the lower basal turn. Three points are apparent from these observations. First, all interventions (including cochleostomy only and round window incision) were associated with reduced OHC counts and SGC densities over the next half cochlear turn apical to the intervention site. Second, electrode array insertion did not cause greater injury to OHCs and SGCs when performed via a cochleostomy. Third, round window insertion caused much more extensive injury to sensorineural structures in the cochlea, and over a greater cochlear distance. Loss of neurosensory structures apical to the site of cochlear implantation (O'Leary et al., 2013; Lee et al., 2013; Eshraghi et al., 2013) or electrode array trauma has been reported previously, and is thought to be associated with oxidative stress followed by hair cell apoptosis (Eshraghi et al., 2013; Hirose and Liberman, 2003). In the lateral cochlear wall, T4 cell densities in the round window insertion group were higher than those for all other treatment groups. No differences were detected between groups for T1, T2, T3 or SV cell densities, or SV area. In summary, these findings suggest that the surgical intervention had little effect upon the lateral cochlear wall. Other types of trauma have been associated with shrinkage of the SV and fibrocyte regions. This has been reported for noise trauma (Hirose and Liberman, 2003; Schuknecht et al., 1974), presbyacusis (Schuknecht et al., 1974; Gratton et al., 1996) and ageing (Hirose and Liberman, 2003; Gratton et al., 1996). Reduced cell densities have been recorded in the SV and T2 and T4 fibrocyte regions in the first few weeks after noise exposure (Mizutari et al., 2011; Hirose and Liberman, 2003; Roberson and Rubel, 1994), but these regions can later be repopulated with cells, most likely due to local proliferation (Mizutari et al., 2011; Roberson and Rubel, 1994; Huber et al., 2010). On the basis of these considerations, it is conceivable that the increased cell density within the T4 region observed here reflects injury to this region associated with the CI surgery. If so, it would appear that round window incision causes less injury to the lateral wall than round
window insertion or the cochleostomy-based interventions. The delayed hearing loss seen in the round window interventions is not readily explained by the histological findings presented here. The finding that cochlear histopathology was more extensive following round window insertion than round window incision, while delayed hearing loss was similar in both groups, suggests that the aetiology of delayed hearing loss is not related to organ of Corti or auditory nerve pathology, nor to the location or extent of the tissue response. Rather, a direct effect upon the round window membrane would seem more likely, as we have suggested above. 4.5. Conclusions This study explored whether the route of electrode array insertion and/or the position of the electrode array within scala tympani influenced hearing and intracochlear pathology. While electrode array trajectory was successfully manipulated in the study, this did not cause appreciable differences in hearing between treatment groups. Round window interventions were associated with a delayed, low-frequency hearing loss. This delayed hearing loss was not related to either electrode array insertion, the tissue reaction in the basal turn of the cochlea, nor the extent of cochlear injury. In fact, cochlear histology suggested that round window incision was the least traumatic intervention, while round window insertion was the most traumatic. However, a thick muscle graft was used to seal the round window in both of these groups, which we speculate changed cochlear mechanics via fibrosis of the round window. Future work to compare cochlear sealants would be useful to further investigate mechanisms of delayed hearing loss. Conflicts of interest The authors have no conflicts of interest to declare. Acknowledgements This work was funded by the Garnett Passe and Rodney Williams Memorial Foundation. The funding body played no role in the design, performance or reporting of this work. The authors thank Gordana Kel (technical assistance), Helen Feng (electrode array manufacture), Michelle Stirling (animal husbandry) and Prue Nielsen (histological processing) for their contributions to the project.
D. Rowe et al. / Hearing Research 333 (2016) 49e57
References Adunka, O.F., Dillon, M.T., Adunka, M.C., King, E.R., Pillsbury, H.C., Buchman, C.A., 2013. Hearing preservation and speech perception outcomes with electricacoustic stimulation after 12 months of listening experience. Laryngoscope 123 (10), 2509e2515. http://dx.doi.org/10.1002/lary.23741. Braun, S., Ye, Q., Radeloff, A., Kiefer, J., Gstoettner, W., Tillein, J., 2011. Protection of inner ear function after cochlear implantation: compound action potential measurements after local application of glucocorticoids in the guinea pig cochlea. ORL J. Otorhinolaryngol. Relat. Spec. 73 (4), 219e228. http://dx.doi.org/ 10.1159/000329791. Burghard, A., Lenarz, T., Kral, A., Paasche, G., 2014. Insertion site and sealing technique affect residual hearing and tissue formation after cochlear implantation. Hear. Res. 312, 21e27. http://dx.doi.org/10.1016/j.heares.2014.02.002. Cervera-Paz, F., Linthicum, F., 2005. Cochlear wall erosion after cochlear implantation. Ann. Otol. Rhinol. Laryngol. 114 (7), 543e546. Choi, C., Oghalai, J., 2005. Predicting the effect of post-implant cochlear fibrosis on residual hearing. Hear. Res. 205 (1e2), 193e200. http://dx.doi.org/10.1016/ j.heares.2005.03.018. Dorman, M.F., Gifford, R., Lewis, K., et al., 2009. Word recognition following implantation of conventional and 10-mm hybrid electrodes. Audiol. Neurootol. 14 (3), 181e189. http://dx.doi.org/10.1159/000171480. Eshraghi, A.A., Polak, M., He, J., Telischi, F.F., Balkany, T.J., van de Water, T.R., 2005. Pattern of hearing loss in a rat model of cochlear implantation trauma. Otol. Neurotol. 26 (3), 442e447 (Discussion 447). Eshraghi, A.A., Gupta, C., van de Water, T.R., et al., 2013. Molecular mechanisms involved in cochlear implantation trauma and the protection of hearing and auditory sensory cells by inhibition of c-Jun-N-terminal kinase signaling. Laryngoscope 123 (Suppl. 1), S1eS14. http://dx.doi.org/10.1002/lary.23902. Gantz, B., Turner, C., Gfeller, K., Lowder, M., 2005. Preservation of hearing in cochlear implant surgery: advantages of combined electrical and acoustical speech processing. Laryngoscope 115 (5), 796e802. http://dx.doi.org/10.1097/ 01.MLG.0000157695.07536.D2. Gfeller, K.E., Olszewski, C., Turner, C., Gantz, B., Oleson, J., 2006. Music perception with cochlear implants and residual hearing. Audiol. Neurotol. 11 (Suppl. 1), 12e15. http://dx.doi.org/10.1159/000095608. Gratton, M.A., Schmiedt, R.A., Schulte, B.A., 1996. Age-related decreases in endocochlear potential are associated with vascular abnormalities in the stria vascularis. Hear. Res. 102 (1e2), 181e190. Greenwood, D.D., 1990. A cochlear frequency-position function for several species e 29 years later. J. Acoust. Soc. Am. 87 (6), 2592e2605. Gstoettner, W., Kiefer, J., Baumgartner, W.-D., Pok, S., Peters, S., Adunka, O., 2004. Hearing preservation in cochlear implantation for electric acoustic stimulation. Acta Otolaryngol. 124 (4), 348e352. Gstoettner, W., Helbig, S., Maier, N., Kiefer, J., Radeloff, A., Adunka, O., 2006. Ipsilateral electric acoustic stimulation of the auditory system: results of long-term hearing preservation. Audiol. Neurotol. 11 (1), 49e56. Hirose, K., Liberman, M.C., 2003. Lateral wall histopathology and endocochlear potential in the noise-damaged mouse cochlea. J. Assoc. Res. Otolaryngol. 4 (3), 339e352. http://dx.doi.org/10.1007/s10162-002-3036-4. Huber, A.M., Hoon, S.J., Sharouz, B., Daniel, B., Albrecht, E., 2010. The influence of a cochlear implant electrode on the mechanical function of the inner ear. Otol. Neurotol. 31 (3), 512e518. http://dx.doi.org/10.1097/MAO.0b013e3181ca372b. Ilberg Von, C., Kiefer, J., Tillein, J., et al., 1999. Electric-acoustic stimulation of the auditory system. New technology for severe hearing loss. ORL J. Otorhinolaryngol. Relat. Spec. 61 (6), 334e340. James, C.J., Fraysse, B., Deguine, O., et al., 2006. Combined electroacoustic
57
stimulation in conventional candidates for cochlear implantation. Audiol. Neurotol. 11 (Suppl. 1), 57e62. http://dx.doi.org/10.1159/000095615. James, D.P., Eastwood, H., Richardson, R.T., O'Leary, S.J., 2008. Effects of round window dexamethasone on residual hearing in a Guinea pig model of cochlear implantation. Audiol. Neurootol. 13 (2), 86e96. http://dx.doi.org/10.1159/ 000111780. Kiefer, J., Pok, M., Adunka, O., et al., 2005. Combined electric and acoustic stimulation of the auditory system: results of a clinical study. Audiol. Neurootol. 10 (3), 134e144. http://dx.doi.org/10.1159/000084023. Lee, J., Ismail, H., Lee, J.H., et al., 2013. Effect of both local and systemically administered dexamethasone on long-term hearing and tissue response in a Guinea pig model of cochlear implantation. Audiol. Neurootol. 18 (6), 392e405. http://dx.doi.org/10.1159/000353582. Merchant, S.N., 2010. Schuknecht's Pathology of the Ear. PMPH, USA. Mizutari, K., Nakagawa, S., Mutai, H., Fujii, M., Ogawa, K., Matsunaga, T., 2011. Latephase recovery in the cochlear lateral wall following severe degeneration by acute energy failure. Brain Res. 1419, 1e11. http://dx.doi.org/10.1016/ j.brainres.2011.08.062. Nadol, J.B., Eddington, D.K., Burgess, B.J., 2008. Foreign body or hypersensitivity granuloma of the inner ear after cochlear implantation: one possible cause of a soft failure? Otol. Neurotol. 29 (8), 1076e1084. http://dx.doi.org/10.1097/ MAO.0b013e31818c33cf. O'Leary, S.J., Monksfield, P., Kel, G., et al., 2013. Relations between cochlear histopathology and hearing loss in experimental cochlear implantation. Hear. Res. 298, 27e35. http://dx.doi.org/10.1016/j.heares.2013.01.012. Richards, S.H., 1981. Congenital absence of the round window treated by cochlear fenestration. Clin. Otolaryngol. 6 (4), 265e269. http://dx.doi.org/10.1111/j.13652273.1981.tb01546.x. Roberson, D.W., Rubel, E.W., 1994. Cell division in the gerbil cochlea after acoustic trauma. Am. J. Otol. 15 (1), 28e34. Schuknecht, H.F., Watanuki, K., Takahashi, T., et al., 1974. Atrophy of the stria vascularis, a common cause for hearing loss. Laryngoscope 84 (10), 1777e1821. http://dx.doi.org/10.1288/00005537-197410000-00012. Seyyedi, M., Nadol Jr., J.B., 2014. Intracochlear inflammatory response to cochlear implant electrodes in humans. Otol. Neurotol. 35 (9), 1545e1551. http:// dx.doi.org/10.1097/MAO.0000000000000540. Shepherd, R.K., Clark, G.M., Black, R.C., 1983. Chronic electrical stimulation of the auditory nerve in cats. Acta Otolaryngol. 95 (s399), 19e31. http://dx.doi.org/ 10.3109/00016488309105589. Souter, M., Eastwood, H., Marovic, P., et al., 2012. Systemic immunity influences hearing preservation in cochlear implantation. Otol. Neurotol. 33 (4), 532e538. http://dx.doi.org/10.1097/MAO.0b013e31824bac44. Turner, C., Reiss, L., Gantz, B., 2008. Combined acoustic and electric hearing: preserving residual acoustic hearing. Hear. Res. 242 (1e2), 164e171. http:// dx.doi.org/10.1016/j.heares.2007.11.008. Wada, H., Sugawara, M., Kobayashi, T., Hozawa, K., Takasaka, T., 1998. Measurement of guinea pig basilar membrane using computer-aided three-dimensional reconstruction system. Hear. Res. 120 (1e2), 1e6. Woodson, E.A., Reiss, L.A.J., Turner, C.W., Gfeller, K., Gantz, B.J., 2010. The hybrid cochlear implant: a review. Adv. Otorhinolaryngol. 67, 125e134. http:// dx.doi.org/10.1159/000262604. Xu, S.A., Shepherd, R.K., Chen, Y., Clark, G.M., 1993. Profound hearing loss in the cat following the single co-administration of kanamycin and ethacrynic acid. Hear. Res. 70 (2), 205e215. XU, J., Shepherd, R.K., Millard, R.E., Clark, G.M., 1997. Chronic electrical stimulation of the auditory nerve at high stimulus rates: a physiological and histopathological study. Hear. Res. 105 (1e2), 1e29.
Contents lists available at ScienceDirect
Hearing Research journal homepage: www.elsevier.com/locate/heares
Research paper
Delayed low frequency hearing loss caused by cochlear implantation interventions via the round window but not cochleostomy David Rowe, Scott Chambers, Amy Hampson, Hayden Eastwood, Luke Campbell, Stephen O'Leary* Otolaryngology, Department of Surgery, The University of Melbourne, 2nd Floor, Peter Howson Wing, Royal Victorian Eye and Ear Hospital, 32 Gisborne St, East Melbourne, VIC 3002, Australia
a r t i c l e i n f o
a b s t r a c t
Article history: Received 20 July 2015 Received in revised form 7 December 2015 Accepted 7 December 2015 Available online 29 December 2015
Cochlear implant recipients show improved speech perception and music appreciation when residual acoustic hearing is combined with the cochlear implant. However, up to one third of patients lose their pre-operative residual hearing weeks to months after implantation, for reasons that are not well understood. This study tested whether this “delayed” hearing loss was influenced by the route of electrode array insertion and/or position of the electrode array within scala tympani in a guinea pig model of cochlear implantation. Five treatment groups were monitored over 12 weeks: (1) round window implant; (2) round window incised with no implant; (3) cochleostomy with medially-oriented implant; (4) cochleostomy with laterally-oriented implant; and (5) cochleostomy with no implant. Hearing was measured at selected time points by the auditory brainstem response. Cochlear condition was assessed histologically, with cochleae three-dimensionally reconstructed to plot electrode paths and estimate tissue response. Electrode array trajectories matched their intended paths. Arrays inserted via the round window were situated nearer to the basilar membrane and organ of Corti over the majority of their intrascalar path compared with arrays inserted via cochleostomy. Round window interventions exhibited delayed, low frequency hearing loss that was not seen after cochleostomy. This hearing loss appeared unrelated to the extent of tissue reaction or injury within scala tympani, although round window insertion was histologically the most traumatic mode of implantation. We speculate that delayed hearing loss was related not to the electrode position as postulated, but rather to the muscle graft used to seal the round window post-intervention, by altering cochlear mechanics via round window fibrosis. © 2015 Elsevier B.V. All rights reserved.
Keywords: Cochlear implant Round window Cochleostomy Hearing loss
1. Introduction Electroacoustic hearing (EAS) has led to improved speech perception and music appreciation in cochlear implant (CI) recipients (Gstoettner et al., 2006; Dorman et al., 2009; Adunka et al., 2013; Gantz et al., 2005; Kiefer et al., 2005; Gfeller et al., 2006). Yet
Abbreviations: EAS, electroacoustic hearing; CI, cochlear implant; ABR, auditory brainstem response; SV, stria vascularis; PTA, pure tone average; LBT, lower basal turn; UBT, upper basal turn; LT2, lower second turn; UT2, upper second turn; IHC, inner hair cell; OHC, outer hair cell; SGC, spiral ganglion cell; OSL, osseous spiral lamina * Corresponding author. E-mail addresses: [email protected] (D. Rowe), [email protected] (S. Chambers), [email protected] (A. Hampson), [email protected]. au (H. Eastwood), [email protected] (L. Campbell), sjoleary@unimelb. edu.au (S. O'Leary). http://dx.doi.org/10.1016/j.heares.2015.12.012 0378-5955/© 2015 Elsevier B.V. All rights reserved.
up to one third of CI recipients experience a progressive loss of residual hearing (Gstoettner et al., 2006; Woodson et al., 2010) which threatens to erode the benefits of EAS. The cause(s) of this hearing loss remain elusive. Delayed hearing loss in humans typically occurs in the months after CI surgery, often following a period of stable hearing. This suggests that delayed hearing loss may be associated with either a chronic or slowly progressing intracochlear pathology. One such candidate is the tissue response to the intracochlear electrode array, which is typified by fibrosis, neoosteogenesis, and an infiltrate of inflammatory cells such as mononuclear leukocytes and histiocytes (Cervera-Paz and Linthicum, 2005; Seyyedi and Nadol, 2014) e findings consistent with this being predominantly a foreign body reaction (Gstoettner et al., 2006; Dorman et al., 2009; Adunka et al., 2013; Gantz et al., 2005; Kiefer et al., 2005; Gfeller et al., 2006; Cervera-Paz and Linthicum, 2005; Nadol et al., 2008; Gstoettner et al., 2004; Ilberg Von
50
D. Rowe et al. / Hearing Research 333 (2016) 49e57
et al., 1999; Turner et al., 2008). Giant cells within a foreign body reaction produce free radicals to degrade the implanted material, and have been implicated in some cases of “soft failure” of CI devices (Gstoettner et al., 2006; Adunka et al., 2013; Seyyedi and Nadol, 2014; Nadol et al., 2008; Choi and Oghalai, 2005). If similar mechanisms occur near the organ of Corti, it is conceivable that the sensory neuroepithelium could be damaged, resulting in delayed hearing loss. Intracochlear fibrosis (or osteoneogenesis) could also potentially influence hearing by interfering with cochlear mechanics (Gstoettner et al., 2006; Woodson et al., 2010; Choi and Oghalai, 2005). Fibrosis within the basal turn of scala tympani is thought to dampen acoustic tuning (i.e. hearing) at the apex of the cochlea, while fibrosis of the basilar membrane is thought to dampen acoustic tuning locally (Cervera-Paz and Linthicum, 2005; Seyyedi and Nadol, 2014; Choi and Oghalai, 2005; O'Leary et al., 2013). Consistent with this, we have observed an association between extensive fibrosis and gradual progression of hearing loss in guinea pigs over the first month after surgery (O'Leary et al., 2013; Shepherd et al., 1983; Xu et al., 1993; James et al., 2008). In light of these considerations, intracochlear fibrosis could conceivably cause delayed hearing loss if its physical characteristics change over time. This study intended to examine whether the trajectory of the electrode array within scala tympani might influence delayed hearing loss. We reasoned that the electrode array trajectory would determine the location and possibly the extent of the foreign body (“tissue”) response, and therefore its effects upon cochlear mechanics or the organ of Corti. The electrode array trajectory was controlled by using the two main surgical approaches to cochlear implantation, namely insertion via the round window (to orient the electrode near to the basilar membrane), or a cochleostomy. To further control electrode array position, we developed a cochleostomy technique that caused the electrode array to traverse scala tympani either more medially or laterally relative to the modiolus. Auditory brainstem responses were recorded at selected time points over a 12 week period, at which time cochleae were harvested then subjected to histopathological analysis.
2. Materials and methods 2.1. Experimental design All experiments were approved by and conducted according to the guidelines of the Animal Research Ethics Committee of the Royal Victorian Eye and Ear Hospital (principal investigator David Rowe; application 11/236AR). The study was performed on 35 Dunkin-Hartley tricolour guinea pigs (350 g). These were randomly allocated to five treatment groups. In two round window treatment groups, the round window membrane was incised, with an electrode array then inserted (round window insertion group) or not inserted (round window incision group) into the scala tympani. The remaining treatment groups all underwent scala tympani cochleostomy, followed in one group by laterally directed insertion of an electrode array (lateral insertion group), or by medially directed insertion of an electrode array in another group (medial insertion group). Auditory brainstem responses (ABRs) were recorded before surgery and 1, 4 and 12 weeks after surgery, after which subjects were euthanised and their cochleae harvested for histological analysis. . All subjects possessed normal hearing (defined here as an ABR threshold to a 100 ms click less than 48 dB peakepeak equivalent) before surgery.
2.2. Auditory brainstem responses Subjects were anaesthetised with a intramuscular injection of mixed ketamine (40 mg/kg) and xylazine (4 mg/kg). The ABR recording system has been described previously (Shepherd et al., 1983; Xu et al., 1993; James et al., 2008; Merchant, 2010). Computer-generated acoustic stimuli (clicks and tone pips lasting 5 ms with 1-ms rise/fall times, with tone pips presented at frequencies of 2, 8, 16, 24 and 32 kHz) were delivered to anaesthetised subjects via a loudspeaker (Richard Allen DT-20, UK) placed 10 cm from the pinna. ABRs were recorded differentially using subcutaneous needle electrodes placed at the vertex and nape. A grounding electrode was placed further caudally in lateral subcutaneous tissue. Ear mould compound (Otoform, Dreve, Germany) was used to attenuate hearing in the contralateral ear. Using this method, the frequency-specific cross-head attenuation (measured from the occluded ear canal of the guinea pig) is 28, 27, 35, 45 and 28 dB at 2, 4, 8, 16 and 32 kHz, respectively. Responses were amplified by a factor of 100,000 (DAM-5A, WeP Instruments Inc., USA) and band-pass filtered (Krohn-Hite 3750, Avon, USA) between 150 Hz and 3 kHz (6 dB/octave). The filter output was fed to a 16- bit analogue-to-digital converter (Tucker Davis Technologies, USA) and sampled at 20 kHz for a period of 10 ms following stimulus onset. Thirty stimuli were presented per second. Responses were averaged over 250 stimulus repetitions. Stimulus intensity was decremented in 5 dB steps from high stimulus levels to sub threshold. Waveforms were exported to a software analysis program (written by Dr. James Fallon, adapted by Prof. Stephen O'Leary using Igor 5.02, Wavemetrics Inc., USA) and analysed by the researcher who was blinded to subjects' identity and treatment group. Threshold was defined as the lowest intensity stimulus that evoked a response of >0.4 mV amplitude in wave III of the ABR (responses of this amplitude reliably exceeded background noise). 2.3. Electrode array The electrode array comprised three platinum rings, each welded to a platinum wire 25 mm in diameter, all housed within a Silastic® carrier (MDX4-4210, Dow Corning Products, USA). The electrode tip was tapered to a diameter of 0.41 mm and had a maximum diameter of 0.45 mm. The three platinum rings, each separated by 0.75 mm, were situated near the tip of the electrode and served as visible markers of insertion depth. The total length of the array was 15 mm. All electrode arrays were sterilised before use. 2.4. Cochlear interventions Guinea pigs were anaesthetised with an intramuscular injection of mixed ketamine (60 mg/kg) and xylazine (4 mg/kg). Local anaesthesia was induced at the target ear with a subcutaneous injection of 2% lignocaine. A curvilinear post-auricular incision was made through skin and subcutaneous tissues to expose the bulla. A #11 blade on a #3 handle rotated in the hand was used to make a bullostomy, through which the round window and basal turn were sighted. Care was taken not to disturb the ossicles or external auditory canal. The cochlea was accessed either by incising the round window membrane with a micro pick (round window groups) or by using a 1 mm diamond bur to drill a cochleostomy in the basal turn adjacent to the round window (cochleostomy groups). In the round window, medial and lateral insertion groups, an electrode array was inserted into the scala tympani until resistance was encountered. The round window or cochleostomy was then sealed with harvested muscle. In the lateral and medial
D. Rowe et al. / Hearing Research 333 (2016) 49e57
insertion groups, the muscle was positioned either inferiorly (medial insertion) or superiorly (lateral insertion) to the electrode array, in order to guide the array towards its desired trajectory in scala tympani. In all insertion groups, the distal end of the electrode array was severed with scissors where it exited the bulla then discarded. The dissected subcutaneous tissues and skin were closed with sutures. Subjects were then administered with a subcutaneous injection of analgesic (Temgesic™, Reckitt Benckiser, Australia) and allowed to recover.
51
according to Wada's adaptation (Choi and Oghalai, 2005; Wada et al., 1998) of Greenwood's equations (Greenwood, 1990). 2.7. Statistics All statistical analyses were performed with IBM SPSS Statistics Version 22 (IBM Corp., USA). Data were variously analysed with Pearson's correlation, chi-square test, Mann Whitney U test and ANOVA, the latter incorporating post-hoc tests and applying a Greenhouse-Geisser correction where required.
2.5. Euthanasia and specimen preparation 3. Results At the conclusion of the 12 week ABR, subjects were euthanised under ketamine/xylazine anaesthesia with an intraperitoneal injection of phenobarbitol. After respiratory arrest, subjects were intracardially perfused with 0.9% saline followed by 10% formalin. Cochleae were harvested and immersed in 10% formalin for at least 48 h. Cochleae were then decalcified over 2e3 weeks in 4% ethylenediaminetetraacetic acid before being embedded in Spurr's resin. Specimens were oriented so that the round window membrane was sectioned in a transverse plane. Sections were collected at 125 mm intervals and stained with hematoxylin and eosin. 2.6. Histological analyses Histological analyses were adapted for the guinea pig from a method described by Merchant (2010). Analyses began at the centre of the modiolus (mid-modiolar), at the section that contained the largest extent of cochlear nerve. To define the scope of analyses, we measured the area of the spiral limbus at the middle of the modiolus within each cochlear turn. We then restricted all other histological analyses to cochlear turns in which the area of the spiral limbus was 25% of that recorded at mid-modiolar. This was done because histological structures become distorted in sections approaching the periphery of the cochlea. Analyses were limited to the basal and second turns of cochleae because, according to the Greenwood equation (Greenwood, 1990), these turns encompass areas responsible for sensing 2, 8, 16, 24 and 32 kHz sounds (matching the frequencies of our pure-tone ABRs). Slides were digitally scanned using a Mirax Scan™ (Zeiss, Germany) slide scanner with a 20 optical objective lens. Pannoramic Viewer™ 1.15.2 (3DHistech, Hungary) was used to count inner and outer hair cells and pillar cells, and to estimate the density of cells in Rosenthal's canal, the stria vascularis (SV), and the T1, T2, T3 and T4 regions of the lateral cochlear wall as described by Mizutari et al. (2011). We also scored the condition of the osseous spiral lamina (OSL; 0 ¼ intact, 1 ¼ inner lamina fractured, 2 ¼ both laminae fractured) and the extent to which the basilar membrane was contacted by tissue reaction (i.e. fibrosis and osteoneogenesis) (0 ¼ basilar membrane not in contact with tissue reaction, 1 ¼ minority of basilar membrane in contact with tissue reaction, 2 ¼ most but not all of basilar membrane in contact with tissue reaction, 3 ¼ all of basilar membrane in contact with tissue reaction). All analyses were performed on two or more sections per cochlear turn. Amira™ 5.4.3 (Visualization Sciences Group, France) was used to reconstruct cochleae in three dimensions (Fig. 1a) in order to calculate the volume of tissue reaction and the electrode array trajectory within the basal turn of implanted cochleae. To determine electrode array trajectory, sections were aligned and splines assigned to the centre of electrode tract (which was visible within tissue reaction), the medial and lateral cochlear walls, the basilar membrane, and the organ of Corti. The resulting co-ordinate data were used to calculate the electrode array trajectory relative to the medial and lateral cochlear walls and the organ of Corti. Additionally, Amira was used to map frequencies onto cochleae
3.1. Electrode array trajectory Electrode array trajectories were plotted relative to the medial and lateral cochlear walls, the basilar membrane and organ of Corti (Fig. 1b, c). The horizontal position of the electrode array was expressed as a proportion of the width of the scala tympani, with medial placement as 0 and lateral placement as 1. Electrode array trajectories matched their intended paths, i.e. arrays oriented either medially or laterally maintained these positions within scala tympani. Round window insertions were the closest to the organ of Corti and basilar membrane over the majority of their intrascalar course. The electrode array tip generally ended near the basilar membrane and organ of Corti in 8e10 kHz region of the cochlea, irrespective of whether arrays were located medially or laterally within scala tympani. 3.2. Auditory brainstem responses Post-surgical threshold shifts for all treatment groups are shown in Fig. 2, with pre-surgery thresholds listed in Table 1. In the round window insertion group, thresholds shifts at 8, 16, 24 and 32 kHz were not significantly different over time. In contrast, threshold shifts at 2 kHz increased significantly over time (F1.084, 6.505 ¼ 13.524, p ¼ 0.008, Greenhouse-Geisser correction ε ¼ 0.542, partial h2 ¼ 0.693), averaging 6.3 ± 3.7 dB at 1 week and 22.9 ± 3.4 dB at 12 weeks. Post-hoc comparisons showed that 2 kHz threshold shifts differed significantly between 1 and 12 weeks (p < 0.001), but not between 1 and 4 weeks after surgery (p ¼ 0.752). The pure tone average (PTA) threshold shift (the average threshold shift across all tested frequencies) for the contralateral ear was 4.1 ± 3.5 dB at 12 weeks. Threshold changes in the round window incision group showed a similar pattern over time to that observed for round window insertions. Threshold shifts at 8, 16, 24 and 32 kHz were not significantly different over time. In contrast, 2 kHz thresholds increased from 1 week (11.1 ± 3.1 dB) to 12 weeks (26.1 ± 3.3 dB) post-operatively. This change was statistically significant (F2, 2 16 ¼ 6.595, p ¼ 0.008, partial h ¼ 0.452). Changes at other frequencies were not significant. The PTA threshold shift for the contralateral ear was 7.1 ± 5.6 dB at 12 weeks. Threshold shifts in the medial insertion group either declined or rose then declined over time across all frequencies measured, although these trends were not statistically significant. Therefore, there was no indication of delayed rise in threshold shifts in this group. The PTA threshold shift for the contralateral ear in the medial insertion group was 1.8 ± 1.5 dB at 12 weeks. Threshold shifts in the lateral insertion group tended to decline over time. The PTA threshold shift for the contralateral ear in the lateral insertion group was 1.4 ± 2.6 dB at 12 weeks. Threshold shifts in the cochleostomy only group did not differ significantly over time. The PTA threshold shift for the contralateral ear in the cochleostomy only group was 1.1 ± 5.4 dB at 12 weeks.
52
D. Rowe et al. / Hearing Research 333 (2016) 49e57
Fig. 1. (a) A cochlea from the round window insertion group three-dimensionally reconstructed from digitally scanned sections, showing the electrode array trajectory (blue) surrounded by tissue reaction (yellow/orange) in the basal turn of scala tympani. Such reconstructions were used to calculate the distance of electrode arrays from (b) the medial and lateral cochlear walls, expressed as a ratio from 0 (most medial) to 1 (most lateral), and (c) the organ of Corti, expressed in mm, in the round window insertion (blue), medial insertion (red) and lateral insertion (green) groups. In (b) and (c), bold lines indicate mean distances while dotted lines bound the standard error of the mean. Array trajectory was measured from the centre of the electrode tract, which was visible within scala tympani tissue reaction (fibrosis and osteoneogenesis). The electrode array had a maximum diameter of 0.45 mm.
An ANOVA showed that threshold shifts at 12 weeks did not differ between interventions at 8, 16, 24 or 32 kHz. However, threshold shifts at 12 weeks for the 2 kHz stimulus were significantly different between treatment groups (F4, 33 ¼ 4.457, p ¼ 0.005). A post-hoc analysis revealed a significant difference between threshold shifts in the round window incision group (26.1 ± 9.9 dB) and the lateral insertion group (7.9 ± 8.6 dB), with a difference of 18.2 dB (95% CI, 2.8e33.7, p ¼ 0.014). 3.3. Histopathology and electrode array trajectory The volume of scala tympani tissue reaction for all treatment groups is shown in Fig. 3, with fibrosis and osteoneogenesis presented separately. The three electrode array insertion groups exhibited significantly greater volumes of tissue response (U ¼ 214, z ¼ 2.810, p ¼ 0.004) than the controls (round window incision and cochleostomy-alone), although 44% of cochleae subjected to cochleostomy-alone contained tissue exceeding 10% of the basal scala tympani volume (this was not seen in any round window incision cochleae). The route of electrode array insertion did not have an effect on the volumes of fibrosis (c (Dorman et al., 2009) (2) ¼ 0.097, p ¼ 0.953) or osteoneogenesis (c2(2) ¼ 0.859, p ¼ 0.651). The volume of new bone was correlated with the total volume of tissue reaction within the cochlea (r ¼ 0.659, p ¼ 0.002). Electrode insertion groups showed a similar degree of basilar membrane fixation in the basal turn (c2(2) ¼ 0.189, p ¼ 0.910). Detailed histological analyses were made at the mid-modiolar plane of the lower basal turn (LBT), upper basal turn (UBT), lower second turn (LT2) and upper second turn (UT2) of each cochlea. Repeated measures ANOVAs were undertaken to determine whether histopathology differed with cochlear location (within subjects) or the type of surgical intervention (between subjects). In the organ of Corti, inner hair cell (IHC) counts differed significantly across cochlear locations (F3, 81 ¼ 3.013, p ¼ 0.035). IHC counts were reduced in the basal turn in all three of the cochleostomy intervention groups, but neither of the round window intervention groups. Between groups, surgical intervention approached significance (p ¼ 0.066), and post-hoc testing indicated that IHC counts following round window insertion were significantly lower than those following cochleostomy only (MD ¼ 0.198, p ¼ 0.025), lateral insertion (MD ¼ 0.238, p ¼ 0.015) and medial insertion (MD ¼ 0.261, p ¼ 0.011) (Fig. 4). Outer hair cell (OHC) counts differed significantly across cochlear locations (F3, 81 ¼ 6.820, p < 0.001), with counts lowest in the lower-basal cochlear turn after
all interventions. Surgical intervention had a significant effect on OHC counts (F4, 27 ¼ 2.892, p ¼ 0.041), with post-hoc testing showing that OHC counts were lower after round window insertion than after round window incision (MD ¼ 1.066, p ¼ 0.007), cochleostomy only (MD ¼ 0.916, p ¼ 0.01), lateral insertion (MD ¼ 0.952, p ¼ 0.015) and medial insertion (MD ¼ 0.938, p ¼ 0.021). Pillar cell counts did not differ across cochlear locations (F1.895,53.006 ¼ 1.382, p ¼ 0.260, ε ¼ 0.632), but were dependent upon surgical intervention (F4, 28 ¼ 9.049, p < 0.001), with post-hoc testing revealing that pillar cell counts were significantly lower after round window insertion than after round window incision (MD ¼ 0.736, p < 0.005), cochleostomy (MD ¼ 0.796, p < 0.005), lateral insertion (MD ¼ 0.713, p < 0.005) and medial insertion (MD ¼ 0.896, p < 0.005). Spiral ganglion cell (SGC) densities were lowest in the lowerbasal cochlear turn, and differed significantly across cochlear locations (F2.306, 64.556 ¼ 12.923, p < 0.001, ε ¼ 0.769). On post-hoc testing, SGC densities were higher following round window incision compared with round window insertion (MD ¼ 106.541, p ¼ 0.039) or cochleostomy only (MD ¼ 126.565, p ¼ 0.008), but did not differ significantly from medial or lateral insertions. In the lateral cochlear wall, T4 cell densities differed significantly with surgical intervention (F4, 28 ¼ 7.412, p < 0.001), and were higher for round window insertion than other approaches [round window incision (MD ¼ 712.485, p < 0.001), medial insertion (MD ¼ 604.082, p ¼ 0.001), lateral insertion (MD ¼ 442.393, p ¼ 0.006) or cochleostomy only (MD ¼ 577.649, p < 0.001)] (Fig. 5). Surgical intervention did not affect T1 cell densities (F4, 27 ¼ 2.348, p ¼ 0.080), T2 cell densities (F4, 28 ¼ 1.273, p ¼ 0.304), T3 cell densities (F4, 28 ¼ 1.411, p ¼ 0.256) or SV cell densities (F4, 28 ¼ 0.918, p ¼ 0.467). 3.4. Site-specific histopathology To explore local relationships between histopathological features, these were correlated within each surveyed cochlear region (the lower-basal turn [LBT], upper basal turn [UBT], lower second turn [LT2] and upper second turn [UT2]). Within the organ of Corti, IHC and OHC counts were positively correlated within most cochlear turns (LBT: r ¼ 0.999, p < 0.001; UBT: r ¼ 0.522, p ¼ 0.002; UT2: r ¼ 0.485, p ¼ 0.004). Similarly, pillar cell counts were positively correlated with both IHC counts (UBT: r ¼ 0.576, p < 0.001; UT2: r ¼ 0.642, p < 0.001) and OHC counts (UBT: r ¼ 0.662, p < 0.001; LT2: r ¼ 0.397, p ¼ 0.022; UT2: r ¼ 0.411, p ¼ 0.018). SGN
D. Rowe et al. / Hearing Research 333 (2016) 49e57
53
densities were positively correlated with either IHC counts (UBT: r ¼ 0.473, p ¼ 0.005) or pillar cell counts (LBT: r ¼ 0.599, p < 0.001; UBT: r ¼ 0.425, p ¼ 0.014) in some turns. In the lateral wall, cell densities within one T-cell region tended to correlate (positively) with densities in other regions (Table 2). This applied particularly for T1, T2, T3 and SV regions. Some significant correlations were observed between the histology of the lateral cochlear wall and that of the organ of Corti. T4 cell densities were negatively correlated with OHC counts in the UBT (r ¼ 0.514, p ¼ 0.002) and with OHC and IHC counts in UT2 (OHC: r ¼ 0.370, p ¼ 0.034; IHC: r ¼ 0.450, p ¼ 0.009). T4 cell densities were also negatively correlated with pillar cell counts in each surveyed turn (LBT: r ¼ 0.475, p ¼ 0.005; UBT: r ¼ 0.532, p ¼ 0.001; LT2: r ¼ 0.529, p ¼ 0.002; UT2: r ¼ 0.556, p ¼ 0.001). T2 cell densities were negatively correlated with SGC densities in the LT2 (r ¼ 0.427, p ¼ 0.013), and SV cell densities were negatively correlated with OHC counts in the UBT (r ¼ 0.351, p ¼ 0.045). OHC counts in the LBT were positively correlated with the magnitude of ABR threshold shifts when the latter were averaged across 8, 16, 24 and 32 kHz (r ¼ 0.545, p ¼ 0.001). OSL fracture in the lower basal turn was positively correlated with the density of multinucleated giant cells (MNGC) within each cochlear turn (LBT: r ¼ 0.599, p ¼ 0.001; UBT: r ¼ 0.379, p ¼ 0.030; LT2: r ¼ 0.379, p ¼ 0.030; UT2: r ¼ 0.390, p ¼ 0.044). 4. Discussion 4.1. Overview of results Medially and laterally-directed electrode array insertions followed their expected trajectories. Round window insertion led to an electrode array placement that was closer to the basilar membrane and more centrally located within the scala tympani relative to the medial and lateral cochlear walls. The electrode array tip ended in a similar place regardless of trajectory, most likely because the scalar diameter decreased, constraining the electrode array path. All interventions, including the control groups, produced similar hearing losses when these were averaged across the cochlea, but the round window interventions were associated with a gradual loss of hearing at 2 kHz. Round window insertion caused the most cochlear trauma, being associated with lower OHC and pillar cell counts within the organ of Corti. Round window incision on the other hand caused the least intracochlear trauma. However, both caused a similar degree of delayed, low frequency hearing loss. 4.2. Round window interventions and hearing
Fig. 2. Auditory brainstem response threshold shifts from before surgery to 1 week, 4 weeks and 12 weeks after surgery (mean ± standard error) at each tested frequency. At 12 weeks, the only significant difference in threshold shifts between treatment groups was at 2 kHz. At this frequency, threshold shifts for the round window incision group were significantly larger than those for the lateral insertion group.
Low frequency (2 kHz) hearing deteriorated over time in both of the round window intervention groups, which suggests that this delayed hearing loss was not associated with the presence of the electrode array within scala tympani. Similarly, delayed loss was not a function of the extent of the intracochlear tissue reaction, because there was minimal fibrosis in scala tympani after round window incision, but more extensive fibrosis after electrode insertion (Fig. 3). The extent of injury to the inner ear was lower following round window incision, as evidenced by the greater survival of sensory cells in the organ of Corti (Fig. 4) and lesser effect upon the lateral cochlear wall following this intervention. This suggests that delayed hearing loss may not be related directly to inner ear injury. Rather, delayed hearing loss appears to be associated with the round window intervention itself. This could potentially be related to incision of the round window membrane, or alternatively to the material (muscle) used to seal the round window after surgery.
54
D. Rowe et al. / Hearing Research 333 (2016) 49e57
Table 1 Pre-surgery auditory brainstem response thresholds (mean ± standard error) for each treatment group. Frequency
Threshold (dB SPL) Round window insertion (n ¼ 7)
2 kHz 8 kHz 16 kHz 24 kHz 32 kHz
51.6 39.1 39.0 43.3 62.7
± ± ± ± ±
2.1 4.7 3.6 2.3 3.5
Round window incision (n ¼ 9) 53.6 43.7 40.7 50.1 65.3
± ± ± ± ±
2.6 3.7 3.2 3.1 4.1
Medial insertion (n ¼ 6) 63.8 35.3 29.0 42.3 62.8
± ± ± ± ±
2.7 3.6 2.9 3.6 3.5
Lateral insertion (n ¼ 7) 57.3 32.0 28.3 46.1 60.6
± ± ± ± ±
2.3 2.7 0.7 1.8 2.6
Cochleostomy only (n ¼ 9) 59.1 41.4 39.0 50.7 65.3
± ± ± ± ±
2.6 3.8 2.0 2.8 4.7
Fig. 3. The proportion of the basal scala tympani occupied by (A) fibrosis and (B) osteoneo-genesis (mean ± standard error). The round window, medial and lateral insertion groups (the three groups in which an electrode array was implanted into the cochlea) had significantly more fibrosis than the round window incision and cochleostomy only groups.
Low frequency hearing loss could potentially occur because cochlear mechanics are altered secondary to physical changes at the round window associated with scar maturation after placement of the muscle plug. Choi predicted that scala tympani damping e which would result from reduced round window membrane compliance e would reduce basilar membrane vibrations in the apex of the cochlea (Choi and Oghalai, 2005; Richards, 1981), consistent with the low frequency hearing loss seen here. Further examples of hearing being affected by round window obstruction can be found in the literature. For example, conductive hearing loss has been found to be secondary to congenital absence of the round window membrane in patients. In these cases, removing the excess bone from the predicted location of the round window membrane revealed a functional membrane with some recovery of hearing (Richards, 1981; Eshraghi et al., 2005). While round window membrane damping may have caused the hearing loss, the reason(s) why this was delayed are less clear. We speculate that this may have been due to a change in the compliance of the scar tissue as it matures over time. Another possible mechanism for hearing loss associated with round window surgery is injury or scarring within the middle ear. This is a possibility because of the close proximity between the
Fig. 4. Counts of inner hair cells (IHCs), outer hair cells (OHCs) and pillar cells (PCs) per mid-modiolar histological section in the two most basal cochlear turns (mean ± standard error). Outer hair cell and pillar cell counts differed significantly among treatment groups, with the round window insertion group exhibiting fewer outer hair cells and pillar cells than all other treatment groups.
ossicles and the round window in the guinea pig. Whilst direct contact between the electrode and the ossicles, or scarring of the ossicular chain was not noted during dissection of the cochleae, the histology did not include the ossicles so it is not possible to entirely exclude this possibility. Previous animal studies have used muscle to seal the round window after surgical intervention, but these did not monitor hearing for sufficient time to detect the delayed hearing loss observed here. A study in the rat introduced cochlear electrode arrays via the round window, withdrew them and then sealed the round window with a muscle plug (Eshraghi et al., 2005; James
D. Rowe et al. / Hearing Research 333 (2016) 49e57
55
et al., 2006). This experiment monitored hearing for 7 days after surgery, compared with 12 weeks in our study. A recent guinea pig study inserted dummy CI electrode arrays via either the round window or a cochleostomy, and employed muscle as a tissue seal, in a manner similar to that described here (Burghard et al., 2014). ABR threshold shifts in that study were of a similar magnitude to ours; the investigators did not observe delayed hearing loss, but their observation window was only 28 days, which again was too short to see the delayed hearing loss observed here. It was noted that the ABR thresholds in the contralateral ear of the round window groups were higher than in the cochleostomy groups by a modest degree (between 3 and 4 dB) at the end of the 12 week observation period. Whether this is of real significance is uncertain, given that this difference was not of statistical significance, but it raises the possibility that there may have been a systemic influence on the hearing induced by the round window surgery. 4.3. Cochleostomy interventions and hearing The most notable characteristic of the cochleostomy interventions was that all groups, including cochleostomy without electrode array insertion, exhibited similar ABR threshold shifts. This was surprising, given that previous studies have revealed minimal hearing loss from a small cochleostomy alone, with greater hearing loss arising following electrode array insertion (Braun et al., 2011). We speculate that our results may have been due to the relatively large cochleostomy performed in our study, which was necessary to manipulate the electrode array towards a medial or lateral trajectory. Larger cochleostomies have been associated with greater hearing loss in a clinical setting (James et al., 2008, 2006; Souter et al., 2012; Lee et al., 2013), with larger cochleostomies considered more likely to involve spiral ligament injury, bone dust entering the cochlea, and perilymphatic fluid leakage. It is also possible that in the present experiment, the more extensive muscle plug required to seal the large cochleostomy may have elicited a greater inflammatory response and affected hearing through this mechanism. Despite our concerns that the cochleostomy may have been the major determinant of the hearing loss, several observations are pertinent to the question of delayed hearing loss. In the first month after surgery, the extent of the hearing loss in the cochleostomy groups was very similar to that for the round window interventions (Fig. 2), and most importantly the threshold shifts at 2 kHz did not differ between groups. This means that the cochleostomy models had the sensitivity to detect delayed hearing loss at 2 kHz, so it is still reasonable to conclude that delayed loss is a characteristic of round window, but not cochleostomy interventions. Secondly, the pathology to the organ of Corti associated with the cochleostomy interventions was in general not as extensive as that for round window insertion, and similar or more extensive than round window incision, supporting the interpretation that intracochlear pathology was not the determinant of delayed loss (James et al., 2008; Souter et al., 2012; Lee et al., 2013; XU et al., 1997). 4.4. Cochlear histopathology and electrode array insertion or trajectory Fig. 5. Cell densities in the stria vascularis (SV) and T1, T2, T3 and T4 regions of the lateral cochlear wall (mean ± standard error). Cell densities in the T2 and T4 regions differed significantly among treatment groups. For the T2 region, the round window insertion group had significantly lower cell densities than all other treatment groups. For the T4 region, the round window insertion group had significantly higher cell densities than the cochleostomy, round window incision and medial insertion groups.
The electrode array insertion groups all had significantly more fibrosis in the basal turn of scala tympani than controls, confirming that the presence of a cochlear electrode array promotes a more extensive tissue response. Round window insertion was associated with lower OHC counts and SGN densities across all cochlea regions than either round window incision or the cochleostomy interventions. Following
56
D. Rowe et al. / Hearing Research 333 (2016) 49e57
Table 2 Pearson's correlations (r values) for cell densities in the lateral cochlear wall, across the four most basal cochlear turns. T1 ¼ type 1 fibrocytes, T2 ¼ type 2 fibrocytes, T3 ¼ type 3 fibrocytes, T4 ¼ type 4 fibrocytes, SV ¼ stria vascularis. LBT ¼ lower basal turn, UBT ¼ upper basal turn, LT2 ¼ lower second turn, UT2 ¼ upper second turn. * Correlation significant at 0.05 level (two-tailed). ** Correlation significant at 0.01 level (two-tailed).
T1 cell density
T2 cell density
T3 cell density
T4 cell density
Cochlear turn
T1 cell density
T2 cell density
T3 cell density
T4 cell density
SV cell density
LBT UBT LT2 UT2 LBT UBT LT2 UT2 LBT UBT LT2 UT2 LBT UBT LT2 UT2
1 1 1 1
0.650** 0.597** 0.716** 0.830** 1 1 1 1
0.517** 0.482** 0.484** 0.694** 0.443** 0.389* 0.574** 0.679** 1 1 1 1
0.218 0.603** 0.296 0.237 0.358* 0.495** 0.197 0.250 0.134 0.369* 0.063 0.139 1 1 1 1
0.189 0.407* 0.187 0.529** 0.064 0.407* 0.238 0.569** 0.206 0.513** 0.546** 0.684** 0.281 0.278 0.066 0.003
round window insertion, electrode array trajectories were closer to the organ of Corti, a factor that has previously been associated with greater OHC loss (Choi and Oghalai, 2005; XU et al., 1997). Injury to the organ of Corti does not seem to be due simply to tissue response encroaching on the basilar membrane (Choi and Oghalai, 2005; O'Leary et al., 2013; Lee et al., 2013), since the extent of sensory cell loss was similar across all electrode insertion groups. In the round window insertion group, OHC counts and SGC densities were reduced in the two most basal cochlear turns. In all other groups, OHC counts were only reduced in the lower basal turn. Three points are apparent from these observations. First, all interventions (including cochleostomy only and round window incision) were associated with reduced OHC counts and SGC densities over the next half cochlear turn apical to the intervention site. Second, electrode array insertion did not cause greater injury to OHCs and SGCs when performed via a cochleostomy. Third, round window insertion caused much more extensive injury to sensorineural structures in the cochlea, and over a greater cochlear distance. Loss of neurosensory structures apical to the site of cochlear implantation (O'Leary et al., 2013; Lee et al., 2013; Eshraghi et al., 2013) or electrode array trauma has been reported previously, and is thought to be associated with oxidative stress followed by hair cell apoptosis (Eshraghi et al., 2013; Hirose and Liberman, 2003). In the lateral cochlear wall, T4 cell densities in the round window insertion group were higher than those for all other treatment groups. No differences were detected between groups for T1, T2, T3 or SV cell densities, or SV area. In summary, these findings suggest that the surgical intervention had little effect upon the lateral cochlear wall. Other types of trauma have been associated with shrinkage of the SV and fibrocyte regions. This has been reported for noise trauma (Hirose and Liberman, 2003; Schuknecht et al., 1974), presbyacusis (Schuknecht et al., 1974; Gratton et al., 1996) and ageing (Hirose and Liberman, 2003; Gratton et al., 1996). Reduced cell densities have been recorded in the SV and T2 and T4 fibrocyte regions in the first few weeks after noise exposure (Mizutari et al., 2011; Hirose and Liberman, 2003; Roberson and Rubel, 1994), but these regions can later be repopulated with cells, most likely due to local proliferation (Mizutari et al., 2011; Roberson and Rubel, 1994; Huber et al., 2010). On the basis of these considerations, it is conceivable that the increased cell density within the T4 region observed here reflects injury to this region associated with the CI surgery. If so, it would appear that round window incision causes less injury to the lateral wall than round
window insertion or the cochleostomy-based interventions. The delayed hearing loss seen in the round window interventions is not readily explained by the histological findings presented here. The finding that cochlear histopathology was more extensive following round window insertion than round window incision, while delayed hearing loss was similar in both groups, suggests that the aetiology of delayed hearing loss is not related to organ of Corti or auditory nerve pathology, nor to the location or extent of the tissue response. Rather, a direct effect upon the round window membrane would seem more likely, as we have suggested above. 4.5. Conclusions This study explored whether the route of electrode array insertion and/or the position of the electrode array within scala tympani influenced hearing and intracochlear pathology. While electrode array trajectory was successfully manipulated in the study, this did not cause appreciable differences in hearing between treatment groups. Round window interventions were associated with a delayed, low-frequency hearing loss. This delayed hearing loss was not related to either electrode array insertion, the tissue reaction in the basal turn of the cochlea, nor the extent of cochlear injury. In fact, cochlear histology suggested that round window incision was the least traumatic intervention, while round window insertion was the most traumatic. However, a thick muscle graft was used to seal the round window in both of these groups, which we speculate changed cochlear mechanics via fibrosis of the round window. Future work to compare cochlear sealants would be useful to further investigate mechanisms of delayed hearing loss. Conflicts of interest The authors have no conflicts of interest to declare. Acknowledgements This work was funded by the Garnett Passe and Rodney Williams Memorial Foundation. The funding body played no role in the design, performance or reporting of this work. The authors thank Gordana Kel (technical assistance), Helen Feng (electrode array manufacture), Michelle Stirling (animal husbandry) and Prue Nielsen (histological processing) for their contributions to the project.
D. Rowe et al. / Hearing Research 333 (2016) 49e57
References Adunka, O.F., Dillon, M.T., Adunka, M.C., King, E.R., Pillsbury, H.C., Buchman, C.A., 2013. Hearing preservation and speech perception outcomes with electricacoustic stimulation after 12 months of listening experience. Laryngoscope 123 (10), 2509e2515. http://dx.doi.org/10.1002/lary.23741. Braun, S., Ye, Q., Radeloff, A., Kiefer, J., Gstoettner, W., Tillein, J., 2011. Protection of inner ear function after cochlear implantation: compound action potential measurements after local application of glucocorticoids in the guinea pig cochlea. ORL J. Otorhinolaryngol. Relat. Spec. 73 (4), 219e228. http://dx.doi.org/ 10.1159/000329791. Burghard, A., Lenarz, T., Kral, A., Paasche, G., 2014. Insertion site and sealing technique affect residual hearing and tissue formation after cochlear implantation. Hear. Res. 312, 21e27. http://dx.doi.org/10.1016/j.heares.2014.02.002. Cervera-Paz, F., Linthicum, F., 2005. Cochlear wall erosion after cochlear implantation. Ann. Otol. Rhinol. Laryngol. 114 (7), 543e546. Choi, C., Oghalai, J., 2005. Predicting the effect of post-implant cochlear fibrosis on residual hearing. Hear. Res. 205 (1e2), 193e200. http://dx.doi.org/10.1016/ j.heares.2005.03.018. Dorman, M.F., Gifford, R., Lewis, K., et al., 2009. Word recognition following implantation of conventional and 10-mm hybrid electrodes. Audiol. Neurootol. 14 (3), 181e189. http://dx.doi.org/10.1159/000171480. Eshraghi, A.A., Polak, M., He, J., Telischi, F.F., Balkany, T.J., van de Water, T.R., 2005. Pattern of hearing loss in a rat model of cochlear implantation trauma. Otol. Neurotol. 26 (3), 442e447 (Discussion 447). Eshraghi, A.A., Gupta, C., van de Water, T.R., et al., 2013. Molecular mechanisms involved in cochlear implantation trauma and the protection of hearing and auditory sensory cells by inhibition of c-Jun-N-terminal kinase signaling. Laryngoscope 123 (Suppl. 1), S1eS14. http://dx.doi.org/10.1002/lary.23902. Gantz, B., Turner, C., Gfeller, K., Lowder, M., 2005. Preservation of hearing in cochlear implant surgery: advantages of combined electrical and acoustical speech processing. Laryngoscope 115 (5), 796e802. http://dx.doi.org/10.1097/ 01.MLG.0000157695.07536.D2. Gfeller, K.E., Olszewski, C., Turner, C., Gantz, B., Oleson, J., 2006. Music perception with cochlear implants and residual hearing. Audiol. Neurotol. 11 (Suppl. 1), 12e15. http://dx.doi.org/10.1159/000095608. Gratton, M.A., Schmiedt, R.A., Schulte, B.A., 1996. Age-related decreases in endocochlear potential are associated with vascular abnormalities in the stria vascularis. Hear. Res. 102 (1e2), 181e190. Greenwood, D.D., 1990. A cochlear frequency-position function for several species e 29 years later. J. Acoust. Soc. Am. 87 (6), 2592e2605. Gstoettner, W., Kiefer, J., Baumgartner, W.-D., Pok, S., Peters, S., Adunka, O., 2004. Hearing preservation in cochlear implantation for electric acoustic stimulation. Acta Otolaryngol. 124 (4), 348e352. Gstoettner, W., Helbig, S., Maier, N., Kiefer, J., Radeloff, A., Adunka, O., 2006. Ipsilateral electric acoustic stimulation of the auditory system: results of long-term hearing preservation. Audiol. Neurotol. 11 (1), 49e56. Hirose, K., Liberman, M.C., 2003. Lateral wall histopathology and endocochlear potential in the noise-damaged mouse cochlea. J. Assoc. Res. Otolaryngol. 4 (3), 339e352. http://dx.doi.org/10.1007/s10162-002-3036-4. Huber, A.M., Hoon, S.J., Sharouz, B., Daniel, B., Albrecht, E., 2010. The influence of a cochlear implant electrode on the mechanical function of the inner ear. Otol. Neurotol. 31 (3), 512e518. http://dx.doi.org/10.1097/MAO.0b013e3181ca372b. Ilberg Von, C., Kiefer, J., Tillein, J., et al., 1999. Electric-acoustic stimulation of the auditory system. New technology for severe hearing loss. ORL J. Otorhinolaryngol. Relat. Spec. 61 (6), 334e340. James, C.J., Fraysse, B., Deguine, O., et al., 2006. Combined electroacoustic
57
stimulation in conventional candidates for cochlear implantation. Audiol. Neurotol. 11 (Suppl. 1), 57e62. http://dx.doi.org/10.1159/000095615. James, D.P., Eastwood, H., Richardson, R.T., O'Leary, S.J., 2008. Effects of round window dexamethasone on residual hearing in a Guinea pig model of cochlear implantation. Audiol. Neurootol. 13 (2), 86e96. http://dx.doi.org/10.1159/ 000111780. Kiefer, J., Pok, M., Adunka, O., et al., 2005. Combined electric and acoustic stimulation of the auditory system: results of a clinical study. Audiol. Neurootol. 10 (3), 134e144. http://dx.doi.org/10.1159/000084023. Lee, J., Ismail, H., Lee, J.H., et al., 2013. Effect of both local and systemically administered dexamethasone on long-term hearing and tissue response in a Guinea pig model of cochlear implantation. Audiol. Neurootol. 18 (6), 392e405. http://dx.doi.org/10.1159/000353582. Merchant, S.N., 2010. Schuknecht's Pathology of the Ear. PMPH, USA. Mizutari, K., Nakagawa, S., Mutai, H., Fujii, M., Ogawa, K., Matsunaga, T., 2011. Latephase recovery in the cochlear lateral wall following severe degeneration by acute energy failure. Brain Res. 1419, 1e11. http://dx.doi.org/10.1016/ j.brainres.2011.08.062. Nadol, J.B., Eddington, D.K., Burgess, B.J., 2008. Foreign body or hypersensitivity granuloma of the inner ear after cochlear implantation: one possible cause of a soft failure? Otol. Neurotol. 29 (8), 1076e1084. http://dx.doi.org/10.1097/ MAO.0b013e31818c33cf. O'Leary, S.J., Monksfield, P., Kel, G., et al., 2013. Relations between cochlear histopathology and hearing loss in experimental cochlear implantation. Hear. Res. 298, 27e35. http://dx.doi.org/10.1016/j.heares.2013.01.012. Richards, S.H., 1981. Congenital absence of the round window treated by cochlear fenestration. Clin. Otolaryngol. 6 (4), 265e269. http://dx.doi.org/10.1111/j.13652273.1981.tb01546.x. Roberson, D.W., Rubel, E.W., 1994. Cell division in the gerbil cochlea after acoustic trauma. Am. J. Otol. 15 (1), 28e34. Schuknecht, H.F., Watanuki, K., Takahashi, T., et al., 1974. Atrophy of the stria vascularis, a common cause for hearing loss. Laryngoscope 84 (10), 1777e1821. http://dx.doi.org/10.1288/00005537-197410000-00012. Seyyedi, M., Nadol Jr., J.B., 2014. Intracochlear inflammatory response to cochlear implant electrodes in humans. Otol. Neurotol. 35 (9), 1545e1551. http:// dx.doi.org/10.1097/MAO.0000000000000540. Shepherd, R.K., Clark, G.M., Black, R.C., 1983. Chronic electrical stimulation of the auditory nerve in cats. Acta Otolaryngol. 95 (s399), 19e31. http://dx.doi.org/ 10.3109/00016488309105589. Souter, M., Eastwood, H., Marovic, P., et al., 2012. Systemic immunity influences hearing preservation in cochlear implantation. Otol. Neurotol. 33 (4), 532e538. http://dx.doi.org/10.1097/MAO.0b013e31824bac44. Turner, C., Reiss, L., Gantz, B., 2008. Combined acoustic and electric hearing: preserving residual acoustic hearing. Hear. Res. 242 (1e2), 164e171. http:// dx.doi.org/10.1016/j.heares.2007.11.008. Wada, H., Sugawara, M., Kobayashi, T., Hozawa, K., Takasaka, T., 1998. Measurement of guinea pig basilar membrane using computer-aided three-dimensional reconstruction system. Hear. Res. 120 (1e2), 1e6. Woodson, E.A., Reiss, L.A.J., Turner, C.W., Gfeller, K., Gantz, B.J., 2010. The hybrid cochlear implant: a review. Adv. Otorhinolaryngol. 67, 125e134. http:// dx.doi.org/10.1159/000262604. Xu, S.A., Shepherd, R.K., Chen, Y., Clark, G.M., 1993. Profound hearing loss in the cat following the single co-administration of kanamycin and ethacrynic acid. Hear. Res. 70 (2), 205e215. XU, J., Shepherd, R.K., Millard, R.E., Clark, G.M., 1997. Chronic electrical stimulation of the auditory nerve at high stimulus rates: a physiological and histopathological study. Hear. Res. 105 (1e2), 1e29.