Conservation of hearing and protection of hair cells in cochlear implant patients' with residual hearing


  • Esperanza Bas,

    1. Cochlear Implant Research Program, University of Miami Ear Institute, Department of Otolaryngology, University of Miami Miller School of Medicine, Miami, Florida
    Search for more papers by this author
  • Christine T. Dinh,

    1. Cochlear Implant Research Program, University of Miami Ear Institute, Department of Otolaryngology, University of Miami Miller School of Medicine, Miami, Florida
    Search for more papers by this author
  • Carolyn Garnham,

    1. MED-EL Hearing Implants Company, Innsbruck, Austria
    Search for more papers by this author
  • Marek Polak,

    1. MED-EL Hearing Implants Company, Innsbruck, Austria
    Search for more papers by this author
  • Thomas R. Van de Water

    Corresponding author
    1. Cochlear Implant Research Program, University of Miami Ear Institute, Department of Otolaryngology, University of Miami Miller School of Medicine, Miami, Florida
    • Cochlear Implant Research Program, University of Miami Ear Institute, 1600 NW 10th Avenue, RMSB 3160, Miami, FL 33136-1015. Fax: 243-5552
    Search for more papers by this author


This review covers the molecular mechanisms involved in hair cell and hearing losses which can result from trauma generated during the process of cochlear implantation and the contributions of both the intrinsic and extrinsic cell death signaling pathways in producing these trauma/inflammation induced losses. Application of soft surgical techniques to conserve hearing and protect auditory sensory cells during the process of cochlear implantation surgery and insertion of the electrode array during the process of cochlear implantation are reviewed and discussed. The role of drug therapy and mode of drug delivery for the conservation of a cochlear implant patient's residual hearing is presented and discussed. Anat Rec, 2012. © 2012 Wiley Periodicals, Inc.


The cochlea contains auditory hair cells (HCs) that are crucial for the conversion of acoustic energy into neural impulses that travel to the auditory pathways of the central nervous system (see van de Water, in press). The apical turn of the cochlea is important for low-frequency hearing, while the basal turn is responsible for high-frequency audition. When there is significant injury and/or loss of the auditory HCs, hearing loss can ensue and amplification of sound may become necessary to improve sound and speech perception. Cochlear implants are an option for patients who no longer receive adequate speech recognition from maximal hearing amplification.

A cochlear implant (CI) has internal and external components. The internal components comprise the receiver-stimulator and the electrode array, while the external components consist of a microphone that receives sound from the environment and a speech processor that converts sound into electric signals that are transmitted to the receiver-stimulator (see Eshraghi et al., in press). The receiver-stimulator decodes this signal and directs current impulses through the electrode array, which stimulates the spiral ganglion neurons of the auditory nerve. As for normal hearing this nerve relays sound to the brain for audition. In order for this to occur, the electrode array must be surgically and strategically inserted into the scala tympani of the cochlea.

Subsets of CI candidates have some residual hearing in the low frequencies which is receptive to acoustic amplification. Preservation of the cochlea's apical auditory function (low frequencies) allows CI patients to benefit simultaneously from both electric stimulation of the spiral ganglion components by the CI itself and acoustic amplification delivered to the low-frequency region of the cochlea with a hearing aid (von Ilberg et al., 1999; von Ilberg et al., 2011). Residual low frequency hearing allows patients with CIs to better perceive fine pitch differences that are important for speech recognition in noisy environments and also for greater appreciation of music (Gantz et al., 2005; Gfeller et al., 2006). Furthermore, preservation of all structures of the hearing cochlea are likely to lead to better sound quality from the CI, since the presence of healthy excitable tissue should in theory lead to better temporal responsiveness and improved tonotopicity or channel separation, giving less masking of speech information and better pitch salience, for example.

There are several factors that can lead to injury of the auditory structures and thereby adversely affect native apical hearing during cochlear implantation. These can include: (1) acoustic and vibratory trauma from drilling; (2) bone particles and blood products displaced into the scala tympani; (3) mechanical trauma to delicate intracochlear structures from the process of electrode insertion; (4) disturbance of the endocochlear potential and cochlear fluid homeostasis; (5) bacterial infection; and (6) a foreign body reaction in response to the electrode array (Soda-Merhy et al., 2008; Friedland and Runge-Samuelson, 2009). To protect remaining low-frequency audition, “soft surgery” techniques were developed and described to limit the degree of injury to the apical portion of the cochlea from implantation (Lehnhardt, 1993; Cohen, 1997; Kiefer et al., 2004; Eshraghi, 2006; Berrettini et al., 2008; Postelmans et al., 2011).

Although some components of the soft surgery technique for cochlear implantation are anecdotal, several of the atraumatic principles have a scientific basis. This article will highlight the molecular mechanisms involved in death of auditory HC following electrode insertion trauma (EIT) and describe the different steps of soft surgery and their impact on HC protection and conservation of residual hearing.


The cochlea contains two distinct fluid compartments: (1) scala vestibuli and scala tympani and (2) scala media (see Van De Water, 2012). The scala vestibuli and scala tympani are connected via the helicotrema and both contain perilymph, which is an extracellular fluid that consists primarily of sodium ions. Therefore, is assumed that scala vestibuli and scala tympani form a single continuous compartment. The scala media contains endolymph, which unlike perilymph, has high potassium content and is more like an intracellular environment. The endocochlear potential (EP) is the diffusion potential or the voltage (∼ 80 mV) of the endolymph relative to that of the perilymph, and this EP is essential for HC function and therefore normal hearing. The EP is responsible for rapid potassium influx through mechanosensitive channels into the stereocilia of the auditory HCs during vibration of the basilar membrane (von Békésy, 1952; Sauer et al., 1999).

Cochlear implantation requires careful insertion of an electrode array into the scala tympani (i.e., the normal route, with a scala vestibuli insertion done only in very rare occasions when scala tympani insertion is not possible). Mechanical trauma from insertion of an electrode array into the scala tympani can lead to oxidative stress and inflammation that can propagate through the tissues of the cochlea and cause a loss of auditory HCs via necrosis and/or programmed cell death (also referred to as “apoptosis”) (Hengartner, 2000). Necrosis, although originally described as an unorganized accidental cell death mechanism, is now considered a regulated process that is initiated by several triggers, such as DNA damage, release of excessive levels of excitotoxins, and ligation of death receptors (DRs; Galluzzi et al., 2011). This form of caspase-independent cell death is linked to activation of receptor interacting protein (RIP) and poly(ADP-ribose) polymerase (Ha and Sydner, 1999; Holler et al., 2000; Zong et al., 2004). On the other hand, apoptosis can be a caspase-dependent form of cell death that is energy-dependent and highly regulated. It is characterized by nuclear condensation and fragmentation, cleavage of chromosomal DNA, and packaging of the dying cell into smaller apoptotic bodies without plasma membrane rupture. These apoptotic bodies are then eliminated by phagocytosis (Edinger and Thompson, 2004).

Cochlear Implantation and HC Loss

During cochlear implantation, the HCs that remain in the area of the apical turn can undergo apoptosis following a variety of trauma-associated stimuli. These stimuli can include: (1) acoustic and vibrational trauma when drilling of bony structures occurs near the cochlea; (2) direct injury to these auditory HCs from mechanical trauma of the electrode array to the apical basilar membrane and sensory cells; (3) displacement of blood and bone particles into the scala tympani during surgery; (4) trauma to the spiral ligament and stria vascularis that contain cells that are important for maintaining the EP; (5) inadvertent insertion of the electrode array through the basilar membrane into the scala media or even into the scala vestibuli, resulting in mixing of the perilymph and endolymph with resultant alterations of the EP; (6) inflammatory response from bacterial infection of the cochlea; and (7) foreign body reaction (Lehnhardt, 1993; Cohen, 1997; Kiefer et al., 2004; Eshraghi, 2006; Berrettini et al., 2008; Postelmans et al., 2011). Of note, osteoneogenesis and fibrosis that also occur within the cochlea following a traumatic insertion can increase the impedance of the CI electrodes, thereby affecting its function.

These trauma-associated stimuli can initiate an inflammatory reaction and oxidative stress within the cochlea that can injure apical auditory HCs, critical for the preservation of residual low-frequency hearing. Tumor necrosis factor alpha (TNFα) is expressed in the cochlea during and following various traumas to the inner ear (Yoshida et al., 1999; Ichimiya et al., 2000). This phenomenon is well documented in response to acoustic and vibrational traumas, bacterial meningitis, cisplatin ototoxicity, and autoimmune-induced hearing loss (Ichimiya et al., 2000; Satoh et al., 2002; Satoh et al., 2003; Aminopour et al., 2005; Zou et al., 2005; Fujioka et al., 2006; van Wijk et al., 2006; So et al., 2007). TNFα is a proinflammatory cytokine that can initiate programmed cell death of auditory HCs (Dinh et al., 2008a, b; Haake et al., 2009), likely through the activation of both the intrinsic (mitochondria) and extrinsic (DR) pathways of apoptosis. Increase of reactive oxygen species (ROS) within the cochlea is also demonstrated following trauma to the inner ear which can promote apoptosis through associated mechanisms. The importance of soft surgery techniques of cochlear implantation for the reduction of these inflammatory and proapoptotic responses within the cochlea is discussed in a later section.

Mechanical disruption of delicate structures of the cochlea can occur from traumatic electrode insertion, which may include penetration through the basilar membrane separating the scalae tympani and media, injury to Reissner's membrane which separates the scala media from the scala vestibuli, and dissection of spiral ligament and stria vascularis tissues. The mechanisms behind this trauma and loss of residual hearing are unclear, but they likely involve alteration of the EP that can result from injury of the spiral ligament and stria vascularis or mixing of perilymph with endolymph when their boundaries are compromised (Takeuchi et al., 2000; Hequembourg and Liberman, 2001; Teubner et al., 2003; Wangemann et al., 2004). Prolonged depolarization of the cell membranes of auditory HCs may lead to the apoptosis of these cells (Cohen-Salmon et al., 2002; Teubner et al., 2003). Excess removal of perilymph during cochlear implantation may produce similar responses in the cochlea.

Intrascalar blood products (and their by-products) may also initiate programmed cell death of HCs, including those located in the area of the apex, through associated release of proinflammatory cytokines (Sercombe et al., 2002). Shifts in the EP can occur through the degradation of erythrocytes and release of their intracellular potassium into the perilymph. Furthermore, ROS that are produced in response to ferrous iron and other hemoglobin byproducts can promote HC loss (Sadrzadeh et al., 1987; van Bergen et al., 1999).

Oxidative Stress and Inflammatory Response

Electrode implantation trauma (EIT) can result in loss of the anatomical and functional integrity of cochlear receptor tissues. Immediately, an inflammatory response is activated to restore the damaged area. The inflammatory response is a complex biological process that involves vascular and cellular components and soluble substances. As illustrated in Fig. 1, immediately after an insertion trauma injury, mast cells and monocytes release vasoactive amines such as histamine and serotonin, inducing local vasodilatation and increased capillary permeability. The local endothelium is activated and expresses surface molecules that promote adherence and migration of leukocytes to the damaged tissue. The mediators of inflammatory response are varied and derive from precursor cells and plasma, which can be classified according to their biochemical properties into: vasoactive amines, vasoactive peptides, cleavage of complement system lipid mediator products, cytokines, chemokines, and proteolytic enzymes. Another group of important molecules in the inflammatory process are neuropeptides, where substance P and calcitonin gene related peptide (CGRP) have proinflammatory effects and are responsible for neurogenic inflammation. Lipid mediators derived from arachidonic acid are produced by activation of phospholipases, which cleave phospholipid constituents of cell membranes generating prostaglandins, leukotrienes, and platelet-activating factor.

Figure 1.

After mechanical trauma induced by electrode insertion, there is local activation of endothelial cells accompanied by an increase of capillarity permeability and vasodilatation that promote mononuclear cells infiltration into the damaged tissue. These mononuclear cells mature in the tissue into macrophages. The fibrocytes and macrophages located in the damaged area release chemokines that will promote the recruitment of inflammatory cells; inflammatory cytokines, such as interleukin (IL)-1β and tumor necrosis factor alpha (TNFα), which will lead to formation of ROS and highly reactive oxygen species (hROS) and an inflammatory response; and growth factors, that is, transforming growth factor-β which can promote a fibroproliferative response, that can result in the formation of fibrous scar tissue around the CI electrode array.

The electrode array following cochlear implantation can also induce a foreign body reaction and thereby a chronic inflammatory insult to the cochlea. In chronic inflammation, tissue characteristically presents an infiltrate composed mainly of mononuclear cells and signs of angiogenesis, and fibrosis (Cruvinel at al., 2010). The fibrocytes and macrophages located in the damaged area can release chemokines, inflammatory cytokines, such as interleukin-1β (IL-1β), TNFα and growth factors such as transforming growth factor-β1 (TGF-β1) (see Bas et al., 2012). A high level of TGF-β1 leads to wound healing via a fibroproliferative response which involves excessive deposition of collagen along with extracellular matrix, and is the cause of fibrous scar tissue around the CI electrode. This scar tissue may affect apical structures important for low-frequency residual hearing, but also increases impedance and affect CI performance (Choi et al., 2005). This type of wound healing comprises inflammation, angiogenesis, migration and proliferation of fibroblasts, and connective tissue remodeling which leads to scarring. There are three isoforms of TGFβ and although all participate in wound healing, TGF-β1 plays a dominant role in the wound repair process associated with scarring while TGF-β2 and TGF-β3 have been shown to play a key role in embryonic development and also in scarless wound healing as noted to occur in fetuses and young neonates. TGF-βs function as regulatory cytokines and the members of the TGF-βs signal response incorporate Smad-dependent or Smad-independent signaling pathways dependent on cell type. TGF-βs can activate several mitogen-activated protein kinases (MAPKs), including extracellular signal-regulated kinases (ERKs), c-Jun N-terminal kinases (JNKs), and p38 mitogen-activated protein kinase (p38 MAPK). TGF-βs also activate the phosphatidylinositol 3-kinase (PI3K) pathway and signaling involving Ras homolog gene family member A (Rho A) (Lin et al., 2005; Li et al., 2006; Town et al., 2008).

IL-1β and TNFα are multifunctional cytokines involved in the propagation of inflammation and apoptosis. Some of the known signal transduction pathways common to these cytokines include coupling to G-proteins, activation of phospholipase A2, calcium mobilization, and ceramide production. TNFα can activate MAPK subfamilies: JNKs, p38 MAPK and ERK (Fig. 2). These cascades of kinases have different functions and can cross-react at several levels. The JNK pathway is involved in the regulation of TNFα-induced gene expression by phosphorylation of transcription factors—mainly cellular Jun transcription factor (c-Jun) and activating transcription factor-2 (ATF-2), leading to increased activity of activator protein 1 transcription factor (AP-1). The p38 MAPK molecule enhances also the function of AP-1 but mainly through other transcription factors such as E twenty-six (ETS)-like transcription factor 1 (Elk-1) or cAMP response element binding transcription factor (CREB), whereas the ERK pathway enhances the function of the nuclear factor kappa B (NF-κB) and can activate an increase in the expression of cellular Myc transcription factor (c-Myc) that can increase a stressed cell's ability to survive (Cohen et al., 2006a)

Figure 2.

Tumor necrosis factor α activates MAPKs subfamilies: JNKs, p38 MAPK, and ERKs. The JNK pathway is involved in the regulation of TNFα-induced gene expression by phosphorylation of transcription factors mainly c-Jun and activating transcription factor-2 (ATF-2), leading to increased activity of activator protein 1 (AP-1). The p38 MAPK enhances also the function of AP-1 but mainly through other transcription factors such as Elk-1 or CREB, whereas the ERK pathway leads to enhancing the function of the NF-κB.

ROS such as: superoxide; hydrogen peroxide; nitric oxide; and hydroxyl radicals, can produce an excessive level of oxidative stress that promotes apoptosis, when the antioxidant potential of the cell's ability for neutralizing ROS is exceeded. Naturally existing enzymes, such as superoxide dismutase, glutathione peroxidase, glutathione reductase, nicotinamide adenine dinucleotide phosphate (NADP) dehydrogenase, and catalase, can prevent oxidative stress by scavenging free radicals. When the cellular antioxidant defenses are exhausted and there is a shift in the cellular redox state, many important cellular events including activation of JNK signaling, transcription factor activation such as NF-kB and AP-1, gene expression (e.g., c-Jun), and apoptosis take place (Lander et al., 1996; Lo et al., 1996; Turner et al., 1998; Wang et al., 1998; Zhuang et al., 2000; Pantano et al., 2003).

TNFα, which has been shown to induce the production of ROS (Woo et al., 2000), is a multi-functional cytokine involved in inflammation. Under many types of cell pathology, this cytokine can bind to its receptor tumor necrosis factor receptor 1 (TNFR1) and initiate downstream pathways that lead to apoptosis through both intrinsic and extrinsic programmed cell death pathways.

Through the DR extrinsic pathway, TNFα activation of TNFR1 leads to the recruitment of tumor necrosis factor receptor type 1-associated death domain (TRADD) and initiation of the caspase cascade (Tartaglia et al., 1995; Hsu et al., 1995). TRADD can recruit other proteins such as TNFα receptor associated factor 2 (TRAF2) and receptor interacting protein (RIP) that leads to activation of the mitogen-activated protein kinase (MAPK)/JNK signaling pathway (Hsu et al., 1996a, b; Lee et al., 1997). JNK signaling can lead to caspase-independent cleavage of BH3-only protein (Bid), which can initiate the mitochondrial intrinsic pathway through cleaved tBid translocation to the mitochondria where it oligomerizes with Bak (Bcl-2 homologous antagonist/killer) or Bax (Bcl-2 associated X protein) to form pores which causes the preferential release of second mitochondria-derived activator of caspases/direct inhibitor of apoptosis protein binding protein with low pI (Smac/DIABLO) (Deng et al., 2003). Activation of JNK can result in phosphorylation of a transcription factor c-Jun that will travel into the nucleus and induce gene expression of proapoptotic proteins, for example, Bax. Bax and Bak activation can then trigger the mitochondria to release cytochrome c (Cyto c) and caspase cascade signaling for apoptosis (Lei et al., 2002). Furthermore, JNK signaling may induce Fas ligand expression, which can bind to the Fas receptor and also promote apoptosis through the intrinsic pathway (Faris et al., 1998a, b).

Apoptosis: Intrinsic Pathway, Extrinsic Pathway, and Caspase-Independent Cell Death

Apoptosis can be initiated by multiple triggers that act through the intrinsic or extrinsic pathways. Although both pathways can independently initiate apoptosis, the extrinsic pathway can communicate with and activate the mitochondrial death signaling cascade through the truncation of a cytoplasm located death protein called Bid (BH3-interacting domain death agonist) forming its active truncated form, that is, tBid (Li et al., 1998). A summary of these DR (extrinsic) and mitochondria (intrinsic) pathways is shown in Fig. 3.

Figure 3.

TNFα activation of TNFR1 leads to the recruitment of TRADD and results in activation of the initiator caspase-8 which can initiate the cleavage of downstream effector caspases such as caspase-3. TRADD can recruit other proteins such as TNFα receptor associated factor 2 (TRAF2) and RIP that leads to activation of the NF-κB, which is translocated into the nucleus and can initiate the transcription of prosurvival genes like IAP, Bcl-XL, Bcl-2. Caspase-8 cleaves Bid, which can translocate into the mitochondria and initiates the mitochondrial intrinsic pathway. A decrease in the membrane potential of the mitochondria due by an increase in Bax levels promotes the release of prodeath proteins from the intermembrane space of mitochondria into the cytoplasm: Cyto c; Smac/DIABLO, Omi/HtrA2, AIF and Endo G. Cyto c recruits Apaf-1 and initiate the formation of the apoptosome, that will activate caspase-9 and downstream caspases such as caspase-3. Smac/DIABLO will also bind to IAPs, thereby release basal inhibition of caspases. Omi/HtrA2 can initiate apoptosis in a caspase-dependent manner by disrupting caspase-IAP interactions, but also acts in a caspase-independent fashion. AIF and Endo G both translocate to the nucleus after being released from the mitochondria and can induce caspase-independent apoptosis.

The intrinsic pathway, aka the mitochondrial death pathway, occurs when an apoptotic stimulus initiates the release of various prodeath proteins from the intermembrane space of mitochondria into the cytoplasm. The most commonly recognized mitochondrial proteins associated with this signaling pathway include: (1) Cyto c; (2) Smac/DIABLO, (3) Omi/HtrA2 (mammalian homolog of bacterial high temperature requirement protein), (4) AIF (apoptosis inducing factor), and (5) endo G (endonuclease G). Cyto c will recruit Apaf-1 (apoptotic peptidase activating factor 1) and initiate formation of an oligomeric complex called an apoptosome, that will activate caspase-9 and downstream caspases for apoptosis to occur (Li et al., 1997; Budihardjo et al., 1999; Rodriguez and Lazebnik, 1999). On the other hand, Smac/DIABLO will bind to inhibitors of apoptosis proteins (IAPs), thereby releasing basal inhibition of caspases (Du et al., 2000; Srinivasula et al., 2000; Verhagen et al., 2000). Omi/HtrA2 is a mitochondrial serine protease that can also initiate apoptosis in a caspase-dependent manner by disrupting caspase-IAP interactions, but it also acts in a caspase-independent fashion through its protease activity (Hegde et al., 2002; Martins et al., 2002, van Loo et al., 2002). Furthermore, AIF and Endo G both translocate to the nucleus of an affected cell after being released from the mitochondria; AIF is a flavoprotein important for chromatin condensation while Endo G is a proapoptotic DNase that cleaves chromatic DNA into nucleosomal fragments (Susan et al., 1999; Li et al., 2001; Candé et al., 2002).

The extrinsic pathway is also referred to as the DR pathway, as it requires death ligand binding and activation of its complementary DR. Caspase-8 is then activated and a death-inducing signaling complex (DISC) is formed, which is important for the activation of caspase-3 and induction of programmed cell death (Nagata, 1999). DRs belong to the tumor necrosis factor receptor (TNFR) superfamily of cell surface receptors. Most of these receptors have a death domain important for recruiting different proteins of the DISC for subsequent activation of caspases. Although TNFR1 and Fas are the most well characterized DRs, other receptors include TNF-related apoptosis inducing ligand receptor 1 (TRAIL-R1 or DR4), TRAIL-R2 (DR5), and TNF-like receptor apoptosis mediating protein (TRAMP or DR3) (Itoh and Nagata, 1993; Tartaglia et al., 1993; Bodmer et al., 1997; Griffith et al., 1999). Activation of these DRs will recruit several intracellular death domain containing adaptors, such as FAS-associated death domain (FADD) and TRADD, which can initiate the caspase cascade (Tartaglia et al., 1993; Chinnaiyan et al., 1995; Hsu et al., 1995).

Caspase activation is an important factor in programmed cell death; however, several caspase-independent pathways have recently been identified (Kawahara et al., 1998; Sperandio et al., 2000; Dawson and Dawson, 2004). As described earlier, AIF and Endo G can initiate apoptosis through their activity in the nuclei of affected cells. Mitochondrial released AIF can induce chromatin condensation and fragmentation of high molecular weight DNA (Susan et al., 1999; Joza et al., 2001; Candé et al., 2002; Candé et al., 2004). In a similar manner, Endo G can lead to fragmentation of DNA, though interaction with exonuclease and DNase I (Li et al., 2001; Widlak et al., 2001). There is some evidence that Omi/HtrA2 can also participate in caspase-independent cell death through its serine protease activity, although the mechanisms and substrates behind this activity are unclear (Hegde et al., 2002; Egger et al., 2003; Blink et al., 2004).

In summary, programmed cell death is a complex phenomenon that can involve activation of intrinsic and extrinsic pathways that may occur in caspase-dependent and caspase-independent manners. Cochlear implantation can deliver and introduce several proinflammatory and proapoptotic stimuli to the auditory HCs of the apical turn of the cochlea that (in turn) impair residual hearing preservation. Prodeath triggers can act through DR complexes, JNK signaling, oxidative stress, and more direct mitochondrial responses. Acute and chronic inflammatory reactions within the cochlea following electrode insertion can contribute to apoptosis and intracochlear fibrosis. Different surgical techniques have been described to reduce the trauma associated with cochlear implantation and preserve low-frequency native audition.


Low-frequency residual hearing in CI patients can be lost following traumatic electrode insertion into the cochlea. The apical portion of the basilar membrane within the cochlea is responsible for hearing in the low frequencies, while the basal turn is primarily involved with audition in the high frequencies. Injury to the apical turn during cochlear implantation can occur from mechanical trauma from the electrode array insertion or changes in the EP. In addition, activation of several proinflammatory or proapoptotic pathways within the cochlea can occur, resulting in intracochlear fibrosis, osteoneogenesis, as well as apoptosis of auditory HCs (Friedland and Runge-Samuelson, 2009).

By preserving apical hearing during implantation, patients can benefit from both electrical stimulation from the CI and amplification of residual low-frequency hearing. This particular subset of patients who have residual low-frequency hearing can utilize their native hearing to distinguish fine pitch differences that the CI itself cannot currently provide. More recently, hearing preservation during cochlear implantation has been extended to include patients with normal or near-to-normal hearing in the low frequencies and severe-to-profound hearing loss in the high frequencies (Skarzynski et al., 2003). Such patients do not utilize amplification of low frequencies after the surgical intervention, but use their natural low-frequency hearing. Many of these patients would not have been considered for cochlear implantation before because their speech recognition scores were either borderline or better than the criterion set for cochlear implantation. However, such patients are beyond the scope of satisfactory treatment by hearing aids alone. Patients with combined electric (from the CI) and acoustic hearing (from a hearing aid or residual hearing only) can benefit from improvement of speech recognition in quiet and noisy environments and music appreciation (von Ilberg et al., 1999; Gantz et al., 2005; Gfeller et al., 2006; Lorenz et al., 2008; von Ilberg et al., 2011).

A soft surgery technique for preservation of native hearing during cochlear implantation was first described by Lehnhardt (1993). Several modifications and additions to this operative technique have been described since then to improve the chances of protecting residual low-frequency hearing (Cohen, 1997; Kiefer et al., 2004; Roberson et al., 2005; Roland et al., 2005; Eshraghi, 2006; Skarzinski, 2007a; Berrettini et al., 2008; Postelmans et al., 2011). Although the reasoning behind some steps of the soft surgery technique is hypothetical, many components are supported by evidence from studies in animals, human temporal bones, patients, and even artificial cochlear models.

Vibration and Acoustic Trauma During Cochlear Implantation

In the soft surgery technique for cochlear implantation, a routine mastoidectomy and drilling of the facial recess is initially performed. Dissection and drilling at this point proceeds with utmost care to avoid acoustic and vibration-related trauma to the inner ear that can result from inadvertent injury of the ossicular chain or overaggressive drilling of the bony overhang of the round window niche, cochleostomy, underlying endosteum and intracochlear structures (Seki et al., 2001; Zou et al., 2001; Pau et al., 2007; Sutinen et al., 2007).

Acoustic trauma from the drill can also be transmitted to the cochlea via the oval and round windows. Apoptosis of outer HCs from noise exposure (Fig. 3) can occur through the release of Cyto c from the stress-damaged mitochondria and through the formation of an apoptosome cause the activation of caspase-9 and then a subsequent downstream activation of caspase-3 (Hu et al., 2002; Nicotera et al., 2003; Henderson et al., 2006). Caspase-independent mechanisms of cell death have also been described in acoustic injury to the auditory HCs through release of Endo G and AIF from oxidative stress injured mitochondria (Yamashita et al., 2004; Han et al., 2006). TNFα that is released following acoustic trauma, can initiate apoptosis of HCs through the extrinsic (i.e., DR) pathway and downstream cell death mechanisms associated with JNK signaling and oxidative stress (Yamane et al., 1995; Lander et al., 1996; Lo et al., 1996; Turner et al., 1998; Wang et al., 1998; Ohlemiller et al., 1999; Ohinata et al., 2000; Pirvola et al., 2000; Zhuang et al., 2000; Hu et al., 2002; Nicotera et al., 2003; Pantano et al., 2003; Wang et al., 2003; Harris et al., 2005; Henderson et al., 2006). Thus, delicate surgical dissection near the facial recess, ossicles, and cochlea is necessary to protect against HC loss by reducing the prodeath responses than can be initiated following acoustic and vibratory trauma.

Injury to the intermediate cells of the stria vascularis, a structure important for maintaining the endocochlear potential, can occur following vibratory trauma to the ossicles (Seki et al., 2001), which can cause alterations in the endocochlear potential and ultimately apoptosis of auditory HCs important for residual hearing (Takeuchi et al., 2000; Teubner et al., 2002; Wangemann et al., 2004). The exact mechanism behind this phenomenon is still unclear. Vibration-induced inner ear damage has also been associated with significant threshold shifts in an animal model. Older animals are more susceptible to this damage than younger animals, and the degree of threshold shift directly correlates with the duration and intensity of the vibratory stimulus (Zou et al., 2001). In addition, high-frequency temporal bone vibration is associated with more severe threshold shift than lower frequencies (Sutinen et al., 2007), suggesting that if drilling occurs near the human cochlea, low drill speeds may prevent more residual hearing loss during cochlear implantation than use of high drill speeds. This vibration-induced threshold shift is likely due to the initiation of a proinflammatory pathway, characterized by expression of TNFα and its receptors TNFR1 and TNFR2, which can activate downstream pathways that lead to apoptosis of auditory HCs (Takeuchi et al., 2000; Seki et al., 2001; Teubner et al., 2002; Wangemann et al., 2004; Zou et al., 2005; Dinh et al., 2008a, b; Haake et al., 2009).

Injury of the stria vascularis has also been demonstrated to occur following exposure to a damaging level of noise (Hirose and Liberman, 2003). Noise trauma can also result in overstimulation of the inner HCs, resulting in release of a large amount of glutamate in the type I fibers of the eighth nerve. This glutamate excitotoxicity causes postsynaptic ion influx and subsequent swelling and rupture of the dendrite terminals of the auditory nerve (Pujol et al., 1993). Loud noise exposure can also cause edema of the stria vascularis, loss of the intermediate cells, and changes in the EP that can cause auditory HC death (Hirose and Liberman, 2003). Furthermore, there is evidence that both the spiral ganglion and cochlear blood flow can also be affected by exposure to acoustic trauma (Pujol et al., 1993; Lamm and Arnold, 2000).

Intracochlear Bone Dust and Blood

Another component of the soft-surgery technique is copious irrigation and removal of bone dust and pâté from the surgical site prior opening and entering into the cochlea. Bleeding from the mastoid and surrounding soft tissues can be minimized using electrocautery and topical vasoconstrictors such as epinephrine-soaked pledgets. Performing these steps prior to the cochleostomy can prevent undesired bone dust that is a product of surgical drilling and blood from entering into the cochlea during electrode insertion, thereby preventing downstream inflammatory reactions within the inner ear that can affect residual hearing.

There are no studies that clearly demonstrate the impact of intrascalar bone dust on residual hearing preservation after cochlear implantation; however, elevation of thresholds can occur if there is significant growth of fibrous tissue or bone in the scala tympani that impedes vibration of the basilar membrane (Clark et al., 1995). Bone particles that remain in the mastoid cavity can initiate growth of solid bone over time (Gstöettner et al., 2000). Thus, bone dust that is displaced into the scala tympani during cochlear implantation can promote intracochlear osteoneogenesis, which can potentially affect preservation of residual hearing. Bone pâté has in the past been utilized as a material to help create a seal at the cochleostomy site following electrode insertion, however this surgical technique has largely been abandoned. McElveen et al. (2006) published a case series of two patients who required revision surgery following electrode extrusion from the cochlea. In both cases, bone pâté was collected from the mastoid cavity and used to seal the cochleostomy site and secure the electrode array. Re-operation in these two patients demonstrated significant obstruction of the basal portion of the scala tympani from osteoneogenesis. Extensive drilling into the new bone formed within the scala tympani was required for reimplantation of an electrode array. Although there is no definitive evidence to suggest that small amounts of intrascalar bone pâté will affect residual hearing preservation, it is believed that large amounts of bone particles, displaced into the cochlea, may potentially initiate osteoneogenesis—and in some cases this can affect the apical portion of the scala tympani.

There is some direct evidence that blood ingress into the scala tympani can adversely affect native auditory function following cochlear implantation. Radeloff et al. (2007) demonstrated temporary and permanent threshold shifts in a guinea pig model after intrascalar injection of autologous blood following cochleostomy. There were more pronounced threshold shifts of the lower frequencies in the study ear, when compared with the control ear, especially at the 6th and 7th week following cochleostomy. This low-frequency threshold shift is likely to be a direct result of the intrascalar blood. Threshold shifts in the high frequencies were seen in both the study and control groups, which is likely the result of surgical trauma or the addition of volume from the injection into the scala tympani. Potential mechanisms that link the presence of intrascalar blood to loss of low-frequency hearing and apoptosis of auditory HCs of the apical turn include: (1) release of cytokines, such as TNFα, IL-1β, and interleukin-6 (IL-6), in response to blood byproducts; (2) production of ROS from hemoglobin byproducts such as ferrous ions; and (3) alteration of the EP due to influx of potassium ions into the perilymph from degradation of erythrocytes (Sadrzadeh et al., 1987; van Bergen et al., 1999; Sercombe et al., 2002; Prunnell et al., 2005; Dinh et al., 2008).

By limiting the amount of intracochlear bone particles and blood products, residual apical auditory function can potentially be preserved. Using these soft surgery techniques, osteoneogenesis, inflammatory cytokines, oxidative stress, and shifts in the endocochlear potential that can affect the apex of the cochlea can be minimized.

Round Window Versus Traditional Cochleostomy Insertion

The remaining steps of the soft surgery technique for cochlear implantation refer to surgical approach to the scala tympani, the use of insertion lubricants, electrode parameters, and depth of electrode array insertion. A round window approach or traditional cochleostomy can be used for electrode insertion into the cochlea (Fig. 4). In the round window approach, drilling of the bony overhang of the round window niche may be necessary to obtain adequate exposure of the membrane. When a traditional cochleostomy is performed, the placement of the cochleostomy is usually anterior–inferior or inferior to the round window membrane. In both techniques, the endosteum of the scala tympani is exposed with a 1-mm diamond burr at low drill speeds with irrigation. Care is taken not to touch the endosteum with the drill.

Figure 4.

A macro-photograph of a right side human temporal bone depicting round window membrane and traditional cochleostomy approaches to the scala tympani of the cochlea. A facial recess approach is performed following routine mastoidectomy. The boundaries of the facial recess are the facial nerve (FN), chorda tympani (CT), and the incus buttress (IB). The round window (RW) and promontory are seen just inferior to the stapes (St). Insertion of the electrode array can be performed through the RW membrane or through a cochleostomy that is performed either inferior (I), anterior–inferior (A–I), anterior (A), or superior (S) to the RW. Soft surgery techniques of cochlear implantation support both RW and traditional cochleostomy approaches that are in a location that are either A–I or A in their anatomical relationship to the RW.

There are advantages and disadvantages to both the round window and traditional cochleostomy approaches. The round window technique has been argued to be less traumatic as it gives you a guarantee of entrance into the scala tympani and does not expose the delicate structures of the cochlea to acoustic and vibration-induced trauma from the drilling of a cochleostomy. However, often times, drilling of the bony overhang of the round window niche is necessary to improve access into the cochlea. Usami et al. (2011) showed that the average drilling time using the round window insertion is about five times lower than the drilling time required for a cochleostomy approach, thus decreasing the average surgical time. In addition, Usami and his colleagues showed that another advantage of using the round window approach is that it preserves vestibular function (i.e., VEMP responses remain within normal limits).

Damage to the modiolus, osseous spiral lamina, or basilar membrane is not commonly seen with round window insertion of many short and standard electrode arrays (Adunka et al., 2004a, b; Briggs et al., 2005; Lenarz et al., 2006). However, perimodiolar CI electrode arrays are associated with increased intracochlear trauma using this surgical approach (Adunka et al., 2006; Souter et al., 2011). With the round window approach, the bony margin anterior–inferior to the round window affects the trajectory of the perimodiolar electrode array, preventing optimal electrode positioning and increasing the risk of trauma to delicate intracochlear structures during implantation (Souter et al., 2011).

Clinically, the round window approach can provide residual hearing preservation if the soft surgery technique for cochlear implantation is used. Skarzynski and colleagues used a round window approach, where a straight electrode array inserted ∼ 360° (insertion depth 18–20 mm) into the scala tympani. Residual low-frequency hearing was preserved in 9 of 10 adult patients for up to 1 year following cochlear implantation (Skarzynski et al., 2007a). In children, full or partial preservation of low-frequency hearing was possible in eight of nine patients with the round window approach (Skarzynski et al., 2007b). In the latest studies, using the straight standard or Flex electrode arrays with the insertion from 20 to 31 mm the hearing preservation rates varies from 81% to 100% (Skarzynski and Lorens, 2010a, b, 25/25; Prentiss et al., 2010, 18/18; Skarzynski and Lorens, 2010a, b, 92/95; Usami et al., 2011, 5/5; Skarzynski et al., 2011, 36/42; Helbig et al., 2011a, 5/5; Helbig et al., 2011b; 13/16), giving a mean hearing preservation rate of 95% (133/139).

In a separate study, a high percentage of patients who received a standard insertion of the contoured electrode array through the round window approach lost residual low-frequency hearing following surgery (75%; 6 of 8 patients). These patients, however, did not receive the soft surgery technique for cochlear implantation (Berrettini et al., 2008).

In the traditional cochleostomy approach, the cochleostomy is usually positioned anterior–inferior or inferior to the round window membrane; however, cochleostomies superior or anterior to the round window have also been described (Fig. 4). There is great variability in the optimal position of the cochleostomy amongst clinicians in soft surgery for preservation of residual hearing (Adunka and Buchman, 2007). An atraumatic insertion with this approach requires a trajectory of insertion that runs parallel to the longitudinal axis of the basal turn of the scala tympani (Briggs et al., 2005). However, to avoid inadvertent injury to important intracochlear structures, such as the modiolus, osseous spiral ligament, and basilar membrane, and to prevent unintended insertion into the scala media and scala vestibuli, the position of the cochleostomy is crucial.

In human temporal bone studies, a cochleostomy that is inferior to the round window membrane is anatomically favorable. Inferior cochleostomies were associated with less intracochlear damage than anterior–inferior or even anterior cochleostomy approaches (Briggs et al., 2005; Adunka et al., 2007). Anterior–inferior cochleostomies have a higher risk of fractures to the osseous spiral lamina and avulsions of the spiral ligament from the lateral wall of the scala tympani adjacent to the round window (Adunka et al., 2007). The likelihood of basilar membrane injury and scala vestibuli insertion is increased in anterior cochleostomies (Briggs et al., 2005). However, three-dimensional modeling of the hook region of the cochlea suggests the anterior–inferior approach allows for direct insertion within the scala tympani while avoiding critical cochlear structures (Li et al., 2007).

Clinically, preservation of low-frequency hearing has been associated with anterior–inferior and inferiorly placed cochleostomies. In a multicenter, prospective trial, Garcia-Ibanez et al. (2009) were able to demonstrate a moderate correlation between the cochleostomy site and residual hearing preservation, with trends toward better threshold preservation with anterior–inferior cochleostomies, followed by inferior, and then anterior cochleostomies. Favorable results with anterior–inferior cochleostomies were also demonstrated in other studies (Berrettini et al., 2008; Gantz et al., 2005). Kiefer et al. (2004) demonstrated at least partial low-frequency hearing preservation in 86% of patients receiving an inferior cochleostomy (12 of 14 patients).

The latest clinical papers show that hearing preservation rates with the straight electrode arrays with the insertion depth of 18–31 mm with cochleostomy approach range from 50% to 100% (Gstoettner et al., 2006, 16/23; Baumgartner et al., 2007, 8/16; Gstoettner et al., 2008, 15/18; Helbig et al., 2011a, 13/13; Helbig et al., 2011b, 2/6; Bruce et al., 2011, 13/13), giving the mean hearing preservation rate of 75% (i.e., 67/89 patients).The role of apoptosis of the auditory HCs following basilar membrane disruption, avulsion of the spiral ligament, and fracture of the osseous spiral ligament following cochlear implantation has not been extensively studied especially in relationship to residual hearing preservation. Potential mechanisms involved with HC loss from traumatic electrode insertion can include: (1) dissection of the spiral ligament and subsequent shifts in the EP; (2) penetration of the basilar membrane resulting in EP changes from mixing of perilymph and endolymph; and (3) direct injury to HCs and the basilar membrane of the apex from insertion of the electrode itself.

Degeneration of the fibrocytes in the spiral ligament can theoretically occur following mechanical injury to or avulsion of the spiral ligament following electrode insertion. Type IV fibrocytes of the spiral ligament are rich in carbonic anhydrase and Na+/K+-ATPase and thus important for ion homeostasis of the cochlear duct. Death of these cells can lead to changes in the EP, resulting in auditory HC death and threshold elevations (Hequembourg and Liberman, 2001). The exact mechanisms behind this cell death are unclear but may involve prolonged depolarization of the cell membrane from alteration of the EP, resulting in activation of caspase-3 or the production of an excessive level of oxidative stress (Cohen-Salmon et al., 2002; Teubner et al., 2003). The association between EP shifts and HC death have been studied in other disease processes (Takeuchi et al., 2000; Teubner et al., 2002; Wangemann et al., 2004), and can be inferred to HC loss from alterations in the EP from penetration of the electrode array through the basilar membrane.

The osseous spiral lamina is a thin bony plate that projects outward from the modiolus, supporting the fibers of the cochlear nerve that travel to each segment of the basilar membrane. Fracture and dislocation of the osseous spiral lamina represents the most severe stage of trauma inflicted in the cochlea from electrode insertion (see Eshraghi et al., in press). It results in discrete loss of spiral ganglion cells to the area involved by the fracture itself. If the osseous spiral lamina is fractured in the apical turn, residual hearing will likely be lost due to injury to the spiral ganglion and its nerve fibers that are crucial for the electrical transmission of low-frequency hearing (Adunka et al., 2010).

Thus, appropriate selection and use of the round window and cochleostomy approaches for insertion of the electrode array may impact the surgeon's success in achieving an atraumatic insertion. The surgical approach that is utilized may decrease the risk of injury to the basilar membrane, spiral ligament, and osseous spiral lamina. Apoptotic responses that can be initiated following traumatic electrode insertion from direct injury to the HCs, oxidative stress, and shifts in EP can be mitigated with use of select surgical approaches to the scala tympani that are compatible with a patient's anatomy (see Rask-Andersen et al., 2012). Trauma to cochlear tissues during insertion of an electrode array can also result in an inflammatory response that can initiate ingress of macrophages, the possibility of initiating a low grade infection and if the inflammatory reaction persists it can impact the vascular supply to the area of damage with resultant hypoperfusion causing additional damage.

Use of Sodium Hyaluronate Gel

In the soft surgery technique, sodium hyaluronate gel may be injected over the endosteum prior to penetration of the endosteum with a micro-lancet. Care is taken to avoid suctioning the perilymph following entry into the cochlea's scala tympani to prevent loss of outer HCs in the upper turns of the cochlea (Garcia-Ibanez et al., 2009; Hara et al., 1990). Gloved hands are then wiped with a sterile wet towel to remove any blood or bone dust. A small sterile drape or towel can be placed over the soft tissues surrounding the surgical site to prevent electrode contact with blood and raw tissues. The electrode array is then carefully inserted through the cochleostomy or round window into the scala tympani. Sodium hyaluronate gel may also be placed on the electrode array as a lubricant to reduce intracochlear friction during insertion.

Sodium hyaluronate gel has been suggested to be both protective and (in one study) detrimental for hearing preservation during the process of cochlear implantation. It has been referred to as hyaluronic acid, hyaluronan, and Healon®. In the soft-surgery technique for cochlear implantation, it is applied over the cochleostomy site prior to penetrating the endosteum, to prevent blood products and bone particles from entering the cochlea and perilymph from escaping (Laszig et al., 2002). It can also act as a lubricant for the electrode array during insertion. Sodium hyaluronate was associated with low insertion forces during electrode insertion when used to fill an artificial model of the human scala tympani, which makes hyaluronate a suitable lubricant for cochlear implantation (Kontorinis et al., 2011). Sodium hyaluronate gel has been studied in several animal models and as part of the soft surgery technique in cochlear implantation in humans.

Anniko et al. (1987) published their study regarding transtympanic injection of hyaluronate gel into the middle ear space and the effects on hearing in rats. Hyaluronate gel was concluded to be safe for use in the middle ear as there were no permanent shifts in hearing thresholds seen at 3 months postapplication. In a guinea pig study performed by Bjurstrom et al. (1987), there were no noticeable differences demonstrated in the amount of outer and inner HC loss between the control ear and the treated ear (sodium hyaluronate gel injection into the middle ear) after 14 days. In a different study, sodium hyaluronate gel was place over the round window membrane in rats, and the round window membrane was punctured using a glass probe (Laurent et al., 1991), which is a technique more representative of soft surgery than prior studies. Mean ABR threshold shifts following puncture of the round window returned to presurgical levels 2 months postoperatively. These results were comparable to another test group which received a round window membrane puncture without sodium hyaluronate gel. Morphology studies of the auditory HCs demonstrate some minor structural changes but no loss of auditory HCs between the two experimental groups. The methods of this particular study are rather similar to the cochleostomy that is performed in the soft-surgery technique of cochlear implantation; therefore its results can be reasonably inferred to the human population.

In contrast, Roland et al. (1995) demonstrated that intracochlear injection of sodium hyaluronate gel can cause severe sensorineural hearing loss in rats as measured by direct round window electrocochleography responses to auditory stimuli. This hearing loss was likely not due to neuronal injury as dendrite and axon histology of the spiral ganglion were preserved. Whether this permanent hearing loss is due to adverse sequelae from mechanical trauma to the auditory HCs from the injection of hyaluronate or any toxicity caused by the sodium hyaluronate gel itself is unclear. Thus, direct intracochlear injection of sodium hyaluronate is not advocated in the soft-surgery technique for cochlear implantation at this time for preservation of residual hearing.

Currently, there are no randomized, controlled trials in humans that demonstrate hyaluronate to protect against or promote loss of residual hearing in soft surgery cochlear implantation. However, conservative use of sodium hyaluronate gel in combination with the soft surgery technique of electrode insertion have been associated with hearing preservation in several retrospective and prospective studies (Skarzynski et al., 2002; Kiefer et al., 2004; Fraysse et al., 2006; Balkany et al., 2006; Garcia-Ibanez et al., 2009). Unfortunately, direct correlation between hyaluronate and residual hearing preservation cannot be adequately assessed from many of the study designs. In a multicenter, prospective trial of 28 patients receiving cochlear implantation with the soft surgery technique, the use of sodium hyaluronate gel as a sealant and lubricant for electrode insertion was associated with preservation of residual hearing, however correlation was weak (Garcia-Ibanez et al., 2009). Angeli demonstrated preservation of sensorineural hearing in patients who received sodium hyaluronate gel application to the oval window following stapedotomy, compared with the control group that did not receive the hyaluronate gel (Angeli, 2006). This study in particular suggests that sodium hyaluronate does not adversely affect hearing outcomes in humans when used in small amounts.

The protective effects of sodium hyaluronate in cochlear implantation are indirect. Utilization of this substance over the endosteum prior to penetration into the scala tympani can prevent displacement of bone and blood products into the cochlea that can affect survival of auditory HCs in the apical turn. By using sodium hyaluronate gel as a lubricant on the electrode array, smooth insertion can potentially prevent mechanical injury to vital structures of the cochlea, such as the basilar membrane, spiral ligament, and osseous spiral lamina, reduce associated apoptotic mechanisms and protect against loss of residual hearing.

Depth of Electrode Insertion

When inserting the electrode array into the cochlea with the soft surgery technique, the electrode insertion is slow and the least pressure for smooth passage is applied. Fascia or muscle is harvested and it may be applied around the insertion site of the electrode to create a seal and protect against egress of perilymph or ingress of air.

The depth of insertion of an electrode array has been shown to influence low-frequency hearing preservation in some studies. In histological studies, insertion beyond the initial encounter of resistance can force the electrode array to penetrate the basilar membrane and fracture the modiolus (Adunka and Kiefer, 2006). The resistance that can occur during electrode insertion can be due to a combination of factors, such as patient cochlear anatomy, electrode properties, surgical approach, and the use of lubricants (Erixon et al., 2009; Gstoettner et al., 1997; Kontorinis et al., 2011). Nevertheless, there are many reports of hearing preservation after insertion of the most flexible MED-EL electrode arrays to their full extent.

In a study using fresh human temporal bone specimens, smooth electrode insertions through a cochleostomy often result in shallower insertion depths but are associated with significantly less intracochlear trauma compared with more forceful and deep insertions (Adunka and Kiefer, 2006). Insertion resistance is most commonly caused by dissection of the spiral ligament, and insertion beyond the point of initial resistance can cause the electrode array to kink within the basal turn of the cochlea causing a fracture of the osseous spiral lamina (Lee et al., 2011; Gstoettner et al., 1997).

Thus, short electrodes arrays were designed and intended shallower insertions of standard arrays were performed in attempts to preserve residual low-frequency audition. For the electrode insertion to reach 360° (18–22 mm insertion depth) in the latest studies at least partial hearing preservation in a large number of adults or children was reported between 97% and 100% (Helbig et al., 2011, 18/18 patients; Skarzynski and Lorenz, 2010a, b, 92/95 patients; Prentiss et al., 2010, 9/9 patients; Skarzynski and Lorens, 2010a, b, 15/15 patients; Gstoettner et al., 2009, 9/9 patients) giving the mean hearing preservation rate of 98% (143/146 patients).

Using the soft surgery technique, hearing preservation with deep insertions; that is, 23–31 mm, of a standard or FlexSoft electrode array is not impossible. While the previous papers on this subject show hearing preservation rates that range from 50% to 81% (Skarzynski et al., 2002; Baumgartner et al., 2007), the latest papers demonstrate hearing preservation rates ranging from 77% to 100% in a large number of child and adult subjects (Skarzynski and Lorens, 2010a, b, 10/10 patients; Pretisis et al., 2010, 9/9 patients; Kleine Punte et al., 2010, 1/1 patients; Usami et al., 2011, 5/5 patients; Skarzynski et al., 2011, 36/42 patients; Helbig et al., 2011b, 17/22 patients; Bruce et al., 2011, 13/13 patients), giving the mean hearing preservation rate of 89% (91/102 patients).

In summary, the depth of insertion of the electrode array can affect residual hearing, and it is suggested that the specific soft surgery insertion techniques are applied to different electrode arrays. If electrode insertion is performed beyond resistance, mechanical injury can occur to the basilar membrane and osseous spiral lamina, which can affect the EP, create oxidative stress, and initiate proapoptotic pathways associated with direct injury to and loss of HCs.

Electrode Parameters

Electrode array characteristics are also crucial for preservation of residual hearing—standard versus perimodiolar, straight versus contoured, flexible versus stiff, electrode diameter and tapering profile, and short versus long. There are several human temporal bone studies that evaluate the different characteristics of each electrode array, however there are mixed results between these studies. Detailed discussion on this topic is beyond the scope of this article. In clinical trials, there were no significant differences in hearing preservation between the straight and perimodiolar electrode styles from one manufacturer (Soda-Merhy et al., 2008). Straight electrodes have been used with round window and cochleostomy insertions with good results for protecting native auditory function in the low frequencies (Kiefer et al., 2004; Gstoettner et al., 2004; Gstoettner et al., 2008; Skarzynski et al., 2007a, b). Contoured electrode insertion through the round window approach has not been evaluated; however, they have been used in cochleostomy approaches with positive results regarding residual hearing preservation (James et al., 2005; Berrettini et al., 2008; Garcia-Ibanez et al., 2009). The appropriate choice of electrode array and surgical approach in the soft surgery technique can prevent the initiation of programmed cell death pathways that are associated with traumatic electrode insertion.


Drugs that target different levels of the apoptosis pathways can be utilized to prevent HC loss in the apical turn and preserve residual low-frequency auditory function. Of these drugs, corticosteroids, JNK inhibitors, and antioxidants have been well studied in animal models for residual hearing preservation following electrode implantation. These drugs and their mechanism of action are summarized in Fig. 5.

Figure 5.

A schematic of the inflammatory and prodeath pathways that can be activated during the process of cochlear implantation with indications where inhibitors, steroids, antioxidants, and soft surgery techniques act to conserve hearing and prevent HC death. Traumatic cochlear implantation can result in apoptosis of auditory HCs of the apical turn of the cochlea important for residual low-frequency hearing. Soft surgery techniques can mitigate prodeath pathways associated with mechanical injury to HCs, acoustic and vibratory trauma, penetration of basilar membrane, dissection of spiral ligament, displacement of blood and bone particles into the cochlea, and activation of the inflammatory response. This can minimize alterations in the endocochlear potential (EP) and decrease the production of ROS that promote oxidative stress. The soft techniques of cochlear implantation can reduce downstream signaling associated with activation of the intrinsic (mitochondrial) and extrinsic (DR) pathways that lead to apoptosis through caspase-independent and caspase-dependent manners. Drug therapy can also promote HC protection and residual hearing preservation during cochlear implantation: (1) antioxidants can reduce oxidative stress; (2) corticosteroids can reduce the inflammatory response as well as activate the glucocorticorticoid receptor (GCR) and nuclear NFκB pathway to promote cell survival; and (3) JNK inhibitors can block downstream proapoptotic responses of JNK signaling.

Local dexamethasone (DXMb) infusion into the scala tympani via osmotic minipump was able to attenuate progressive hearing loss associated with electrode implantation trauma (Vivero et al., 2008). The mechanisms behind DXMb protection against auditory HC loss from TNFα-induced programmed cell death incorporates activation of nuclear factor kappa B (NFκB) and up regulation of prosurvival Bcl-2 and Bcl-xl gene expression (Dinh et al., 2008a, b; Haake et al., 2009; Dinh et al., 2010). Prodeath Bax and proinflammatory TNFR1 gene expression associated with TNFα-induced damage were also mitigated with treatment with DXMb. It is believed that corticosteroids can protect against death of HCs through inhibition of proinflammatory and proapoptotic triggers that initiate intrinsic and extrinsic cell death and prevents recruitment of additional macrophages to the site of injury and adjacent sites. In guinea pig studies, postelectrode insertion compound action potential (CAP) and auditory brainstem response (ABR) threshold shifts returned to preoperative levels by 90 days through treatment with DXMb, eluted from silicone embedded in the electrode array (van de Water et al., unpublished data). The results of another drug-eluting electrode study (Eshraghi et al., 2011) has shown that CI electrode blanks either uncoated or coated with a biopolymer (i.e., styrene–isobutylene–styrene; SIBS) inserted into a guinea pig cochlea via a cochleostomy and left for 1 month caused a >25 dB SPL permanent hearing loss in response to a 16 kHz pure tone stimuli and a loss of auditory HCs from the electrode insertion traumatized cochlea (Fig. 6B–D). In contrast, guinea pigs that had their scala tympani implanted with an CI electrode coated with SIBS + DXMb for a period of 1 month experienced an initial hearing loss that was equivalent to that of the uncoated and SIBS only coated CI electrode blanks on the day of electrode insertion trauma (EIT) (i.e., day 1) but fully recovered their hearing thresholds to a presurgical baseline levels (i.e., day 0) by 1 month post-EIT and had no loss of auditory HCs at 1 month post-EIT (Fig. 6A, E). The results of this study (Fig. 6) show that CI electrode eluted DXMb protected both the hearing and the HCs in a animal model of EIT-induced hearing loss (Eshraghi et al., 2011) as effectively as direct infusion of dexamethasone into the scala tympani of an EIT-induced hearing loss animal model (Vivero et al., 2008). Reversal of the CAP threshold shift that occurs following cochleostomy in the guinea pig occurred with topical corticosteroid treatment, when compared with the control groups (Keifer et al., 2007; Ye et al., 2007; Braun et al., 2011). In addition, Eastwood et al. (2010a) were able protect hearing of the basal and second turn of the cochlea in a guinea pig model following DXM delivery through the round window. Intravenous administration of corticosteroids at high doses can also mitigate postimplantation associated hearing loss in an animal model (Connolly et al., 2011). Corticosteroids can prevent apoptosis of auditory HCs by preventing downstream effects of inflammatory cytokine expression (especially TNFα and IL-1β) and protection against the production excessive levels of ROS (see Bas et al., 2012).

Figure 6.

Dexamethasone eluted from a CI electrode array Protects against electrode insertion trauma-induced hearing and HC losses in an animal model of electrode insertion trauma (EIT)-induced hearing loss. (AC) Plots displaying individual and mean auditory evoked brainstem response (ABR) threshold values from animals implanted with: (A) SIBS + dexamethasone base (DXMb) coated electrodes; (B) uncoated electrodes; and (C) SIBS-only coated electrodes in response to 16 kHz pure tone stimuli. Dots represent individual ABR threshold values and solid blue lines plot the mean values (n = 8 animals/group). The plots in (B) and (C) show a large increase (i.e., >25 dB SPL) in ABR threshold values over presurgery baseline values beginning on day of surgery (i.e., day 1) that persists for 1 month post-EIT representing a permanent hearing loss. The plot in A shows the same initial large increase in ABR threshold (i.e., >25 dB SPL) on day of surgery and EIT (i.e., day 1) as was present in the uncoated (B) and the SIBS-only (C) coated electrode animals, but this initial increase in ABRf threshold did not persist and instead returned to near presurgery levels by day 30 post-EIT with no permanent hearing loss. (D, E) Microphotographs of the organ of Corti specimens from the lower middle turn areas of experimental animals stained with FITC-phalloidin to stain f-actin for visualization of the auditory HCs stereociliary bundles and cuticular plates can be identified. (D) A SIBS-only electrode animal organ of Corti and (E) a SIBS + DXMb electrode animal organ of Corti. The images of the SIBS-only organ of Corti specimen (D) shows extensive loss of the auditory HCs while in comparison the image of the organ of Corti specimen dissected from a SIBS + DXMb cochlea (E) shows a normal pattern of intact auditory HCs. Bar in D = 50 μm for both (D) and (E).

Inhibitors of JNK signaling have also demonstrated effectiveness in promoting HC survival and residual low-frequency hearing preservation after electrode insertion trauma. DJNK inhibitor-1 (DJNKI-1) treatment with an osmotic minipump in a guinea pig model was able to substantially attenuate ABR threshold shifts and changes in distortion product otoacoustic emissions (DPOAE) associated with immediate and progressive auditory dysfunction following EIT (Eshraghi et al., 2006, 2007). DJNKI-1 (also known as AM-111) and SP600125 are peptides that bind to the three JNK isoforms and block downstream activity associated with c-Jun phosphorylation and programmed cell death (Bonny et al., 2001; Bennett et al., 2001; Mardersetein et al., 2003; Minutoli et al., 2004; Sharma et al., 2005; Assi et al., 2006). Furthermore, JNK-inhibitors have protective effects against hearing loss associated with acute labyrinthitis, aminoglycosides, and acoustic trauma in an animal model (Wang et al., 2003; Barkdull et al., 2007). The mechanisms behind JNK inhibition and HC protection likely involve decreased phosphorylation of transcription factor c-Jun and reduced Bax and TNFR1 gene expression (van de Water et al., unpublished data).

Lastly, there is an increasing role for antioxidants in the protection of residual hearing loss as there is more evidence that links high levels of oxidative stress to programmed cell death of remaining HCs. Exogeneous antioxidants such as: sodium thiosulfate (STS); D-methionine (D-met); N-acetylcysteine (NAC); glutathione; ascorbic acid; and tocopherol (Kannan and Jain, 2000) have been shown in various studies to decrease the oxidative stress burden on cochlear receptors to prevent hearing loss. However, none of these studies address hearing loss associated with EIT and cochlear implantation except for Eastwood et al. (2010b) who were able to demonstrate benefit with NAC treatment. This benefit only pertained to the high frequencies of the basal turn and did not extend significantly to low-frequency hearing located in the apical section of the cochlea. Unfortunately, delivery of NAC to the round window prior to implantation caused a slight increase in hearing thresholds and greater amounts of osteoneogenesis, which may preclude its use locally in the protection of residual hearing. Antioxidants have shown some degree of benefit against cisplatin, carboplatin, aminoglycoside, and noise-induced trauma to the inner ear (Kopke et al., 2000; Lockwood et al., 2000; Doolittle et al., 2001; Duan et al., 2004; Wimmer et al., 2004; Kopke et al., 2005; Kramer et al., 2006; Kopke et al., 2007; Samson et al., 2008).

As demonstrated in various animal models and human studies, drugs can be delivered to the injured cochlea by direct perilymphatic perfusion, diffusion through the round window membrane, and intravenous dissemination (Chang et al., 2009; Barriat et al., 2010; Staecker et al., 2010). Another novel mode of drug delivery following cochlear implantation is the drug-eluting electrode array (Dinh et al., 2008a; Farahmand-Ghavi et al., 2010; Eshraghi et al., 2011). Micronized dexamethasone eluted from a biopolymer, SIBS, was first reported to protect HCs from the ototoxic effect of an inflammatory process associated TNFα in organ of Corti explants (Dinh et al., 2008). Hearing thresholds and HC preservation were demonstrated in a guinea pig model of EIT treated with elution of DXM from DXM/SIBS coated electrodes (Eshraghi et al., 2011). Furthermore, Farahmand-Ghavi et al. (2010) demonstrated the effectiveness of micronized DXM embedded in silicone elastomer as a method of drug release from a CI electrode array in preservation of hearing. DXM elution from silicone has potential to protect against threshold shifts associated with cochlear implantation long term (Jolly et al., 2011; Van De Water et al., unpublished data). Delivery in this manner has many potential advantages over other methods. The drug dosage can be well controlled, the release pattern can be made stable over time for time periods from weeks to years, if required, and high concentrations in the region of the cochleostomy site can be avoided. Injury of the organ of Corti that houses the auditory HCs can impair release of neutrophins, that is, brain-derived neutrophic factor and neurotrophin-3, which are essential for spiral ganglion nerve survival and attachment (Staecker et al., 1996; Evans et al., 2009; Green et al., 2012; Budenz et al., 2012). Guinea pig ears implanted with electrodes coated in polypyrrole/para-toluene sulfonate containing neurotrphin-3 (Ppy/pTS/NT3) showed a higher density of surviving spiral ganglion neurons compared with controls. In addition, the electrodes did not exacerbate fibrous tissue formation or affect electrode impedance. Although this drug does not particularly affect residual hearing loss, it may improve auditory perception of both the low and high frequencies from CI electrical stimulation. Lastly, stem cell implantation has emerged as a potential modality for augmentation of severely degenerated spiral ganglion neurons that can may transform the role and outcomes of cochlear implantation in patients with unserviceable hearing loss and poorly degenerated auditory nerve (Li et al., 2009; Staecker et al., 2010).


Preservation of residual low-frequency hearing during cochlear implantation can improve speech recognition in noisy environments and music appreciation in a select group of candidates that have some HC function in the apical turn region of the cochlea. Trauma to the cochlea during surgical placement of the electrode can induce various proinflammatory and prodeath stimuli, which can initiate the intrinsic and extrinsic pathways of apoptosis in affected HCs. Programmed cell death of these HCs can occur by caspase-dependent and caspase-independent modalities. The soft techniques of cochlear implantation can mitigate the expression of the triggers and the consequences of traumatic electrode insertion and thereby protect against apoptosis of auditory HCs in the apical turn of the cochlea responsible for residual hearing. Various drug therapies have also been implemented during cochlear implantation to preserve low-frequency audition. These drugs can be delivered to the injured cochlea in several manners. Although residual hearing preservation through soft surgery techniques and drug delivery is a trend in cochlear implantation and HC survival, treatments directed toward spiral ganglion neuron and peripheral process protection and regeneration may present new facets in this arena. One example of this would be that preservation of apical dendrites can be important because these apical neural elements can be stimulated with long pulse durations, resulting in specific pitch percepts. Atraumatic, deep insertion of an implant's electrode array is important for all CI patients, even those without serviceable hearing.