The ototoxicity of gentamicin primarily targets the vestibular system (Ahmed et al. 2012), and this effect can be used when treating debilitating Menière’s disease through intratympanic injections, to achieve unilateral vestibular deafferentation (uVD). The mechanism for vestibular damage is through generation of free radicals that destroy the hair cells of the vestibular organ (Forge & Schacht 2000). Although intratympanic injections with gentamicin are considered to affect cochlear function to a lesser extent, most physicians would advocate a cautious dosage and consider the hearing function before treatment (Cohen-Kerem et al. 2004).
Another situation where intratympanic treatment with gentamicin might be considered is before vestibular schwannoma surgery (Magnusson et al. 2011). The vestibular PREHAB (prehabilitation) protocol was developed to ensure a better rehabilitation, since many, if not most, patients subjected to vestibular schwannoma surgery have to cope with both a sudden loss of remaining vestibular function, as well as with the aftermath of neurosurgery (Magnusson et al. 2009). The protocol encompasses treatment of patients that have measurable vestibular function before surgery with intratympanic gentamicin injections, while performing daily vestibular exercises before and after treatment (Magnusson et al. 2009). Through this procedure, the sensory trauma is separated from the surgical trauma, making it possible for vestibular compensation to ensue gradually and unencumbered by any depression of central nervous function that might result from surgery. Beneficial aspects on postural stability to withstand postural perturbations have been reported after using the PREHAB rehabilitation protocol (Tjernström et al. 2009), supporting this hypothesis.
The indications and aims for schwannoma surgery have changed during the past decades toward hearing preservation (Kari & Friedman 2012). In this perspective, it is important to determine whether pretreatment of schwannoma patients with gentamicin is as safe regarding cochlear damage as when treating Menière’s disease, but at the same time as effective in producing uVD. Patients with small medially located schwannomas are most prone to have close to perfect hearing before surgery and are also the best candidates for hearing preservation surgery (Kari & Friedman 2012). However, those patients probably also have the most vestibular function and would theoretically benefit from the vestibular PREHAB protocol. The aim was thus to examine the outcome in terms of hearing preservation and vestibular deafferentation in schwannoma patients subjected to the PREHAB protocol.
Materials and Methods
Suitable patients were found by a retrospective consecutive survey of medical charts of all patients subjected to the vestibular PREHAB protocol since the start in year 2004 to 2011. In 17 patients, audiometric data could be retrieved before and after gentamicin installations (13 with word recognition score before and after treatment). Fifteen out of these 17 had performed complete vestibular testing after gentamicin treatment. The time between gentamicin treatment and functional evaluation ranged from 4 to 7 weeks, with a median of 6 weeks.
The patients’ age ranged between 21 and 66 years of age (mean 48.8), 9 females and 8 males. Six patients were scheduled for translabyrinthine and 11 for retrosigmoidal surgical approach. All patients received between 2 and 4 (mean 2.7) intratympanic installations of buffered gentamicin 30 mg/mL (Table 1). One of the 2 patients that were treated with 4 injections had an initial treatment of 2 injections; however, the effect was deemed to be insufficient in producing uVD leading to further two injections. The decision to treat with 2, 3, or 4 injections did not depend upon the level of vestibular function, but rather on that the protocol was instigated at a time when all surgical procedures were through the labyrinth, making consideration for hearing preservation obsolete. Furthermore, some patients lived far away from our hospital, which might have induced us to “over treat” to achieve vestibular deafferentation. The method to anaesthetize the tympanic membrane was either performed by the use of topical phenol (carbolic acid) or by topical EMLA® (ointment containing lidocaine and prilocaine).
Informed oral consent was obtained from all patients and the treatment was performed in accordance with the World Medical Association’s Helsinki Declaration, and the local ethical committee approved the research study.
Hearing levels were assessed with pure tones, and an average (PTA) was calculated at 500, 1000, 2000, and 3000 Hz (Gurgel et al. 2012). Hearing levels were examined through air conduction except for 2 patients. In those 2 patients, bone-conducted hearing levels were deemed more accurate because they developed perforations after the intratympanic injections. Speech recognition was tested using two-syllabic words at a level comfortable for the patient. The hearing was further classified according to guidelines from the American Association of Otolaryngology-Head and Neck Surgery (American Academy of Otolaryngology-Head and Neck Surgery Foundation, INC” 1995).
The vestibular system was evaluated through video-recorded head-impulse test (HIT) of all three canals of each ear, bithermal calorics, vestibular-evoked myogenic potential measured on the sternocleidomastoid muscle (cVEMP), and subjective visual horizontal and vertical (SVH-V). With regard to tumor compression on vestibular nerve function, the test of the anterior and lateral HIT, calorics and SVH-V assess the function of the superior branch of the vestibular nerve and the posterior HIT and cVEMP assess the function of the inferior branch.
HITs evaluate the function of each semicircular canal (SCC) to high-frequency motion (Magnusson et al. 2002; Karlberg et al. 2004). There is a risk that s.c. covert eye saccades, that mimic a normal vestibulo-ocular reflex, were undetected with this method, and consequently a pre-existing vestibular deafferentation could have been misdiagnosed (Tjernström et al. 2012). Commercial video-HIT systems were not available at the time when the PREHAB protocol was initiated and could not be included as parameters in the present study. The “uncover” maneuver for clinical detection of uVD (uncovering an overt saccade) had also not been not invented (Tjernström et al. 2012). However as Table 2 indicates, the vestibular function evaluation did not rely solely on one normal functional outcome. Bithermal calorics detect slow-frequency function of the lateral SCC (Schmid-Priscoveanu et al. 2001). A caloric response not exceeding 7.5 degrees/s (best response of either 30 sec warm water [44°C] or 30 sec cold water [30°C] irrigation) was considered to reflect an absent caloric response (Böhmer & Rickenmann 1995). Two patients suffered from perforation of the tympanic membrane after intratympanic treatment, and as a consequence, those 2 were tested with air calorics after treatment.
With cVEMP, it is possible to assess the function of the sacculus in inhibiting vestibulospinal function (Manzari et al. 2010). The response was assessed from a normalized stimulus. If the p1–n1 amplitude recorded was below 40 μV or either of the p1 and n1 reactions could not be detected, we deemed the function as absent. Thus, to be considered uVD, a patient should have pathological HIT of all three SCCs, have a caloric response that does not exceed 7.5 degrees/s, and have no cVEMP response.
uVD causes less or no input from the utriculus which leads to ocular torsion resulting in a tilted subjective appreciation of the gravitational forces, leading to a leaning of the SVH-V toward the side of the lesion (Hafstrom et al. 2006). During the spatial orientation test, the subjects sat upright in a dark room with their heads fixed against a neck rest. A 15-cm-long and 2-mm-wide softly lit green rod was projected on a wall of 1.5 m in front of them. The rod was rotated by the subject with a remote control four times to the perceived gravitational horizontal (the SVH) and four times to the perceived gravitational vertical (the SVV). Thereafter, the perceived spatial orientation error was determined by calculating the average deviations made from perfect horizontal and vertical during the assessments. A tilt of the test by 3 or more degrees is considered to be pathological (Hafstrom et al. 2006). The test, however, does not necessarily assess the function of the utriculus, but rather if a functional change (decrease) happened in the near past or if there is tumor compression on the brainstem/pons. In the PREHAB protocol, a clear change of SVH-V toward the side of the lesion was considered to be a token that the intratympanic injections had an effect.
Hearing thresholds before and after gentamicin therapy was analyzed using the multivariate method repeated measures general linear model (GLM) analysis of variance (ANOVA) test including main factors: treatment (before, after; df 1) and hearing frequency (500, 1000, 2000, 3000, 4000, 6000, and 8000; df 6; Altman 1991). In the GLM analysis, p values <0.05 were considered statistically significant. With GLM ANOVA, the general gentamicin treatment effects together with its impact on individual frequency thresholds can be analyzed. Post hoc analysis included Wilcoxon exact (two tailed) test on individual hearing levels (both PTA and at different frequencies) before and after gentamicin treatment. The Wilcoxon exact (two tailed) test was also used when comparing caloric reaction and SVH-V before and after treatment.
The correlations were performed with Spearman two-tailed test.
In Figure 1 and Table 1, the pure-tone hearing levels in the 17 patients before and after gentamicin treatment are demonstrated. According to GLM ANOVA, the hearing decreased with gentamicin treatment (p < 0.001, F = 24.1), and the hearing loss was most pronounced in the higher frequencies (p < 0.001, F = 46.6; Fig. 1). Moreover, also the average PTA value was significantly lower after gentamicin therapy compared with before therapy (p = 0.004). For 12 of the patients, the PTA did not differ between the two measurements with more than 10 dB (Table 1). Two patients suffered isolated high-frequency loss. Two patients had an overall hearing loss of >10 dB accompanied with a loss of speech recognition (Table 1). The quantitatively largest hearing loss was seen in a patient that had no speech recognition before gentamicin treatment. Overall the PTA was 33.8 ± 17.0 dB before and 40.8 ± 23.4 dB after gentamicin (average PTA loss of 7.1 ± 8.5). There was no correlation between tumor size and PTA before (p = 0.492, R = −0.179) or after gentamicin treatment (p = 0.544, R = −0.158).
Speech recognition was determined in 13 patients (Table 1). The ability to recognize spoken words diminished with an average of 10% at 2.3 dB louder volume.
For nine out of the 13 patients, the hearing classification did not change as a result of the gentamicin treatment (Table 1). Individual hearing of 2 patients before and after gentamicin is presented in Figure 2.
The vestibular test data before and after gentamicin installations are demonstrated in Table 2 and Figures 3 and 4. All patients had at least one or more fully functioning SCC, evaluated by the HIT before gentamicin treatment. All patients displayed pathological HITs in all SCCs 6 weeks after gentamicin injection. Eleven patients had intact cVEMP before gentamicin, although one patient did not perform the test (Table 2). Fifteen had no evoked cVEMP after gentamicin and the remaining 2 patients were not tested after treatment.
In Figure 3, the best response of bithermal calorics are shown before and after gentamicin treatment. There was a significant reduction of the caloric response by 64% (p < 0.001). Six patients were deemed to have no caloric response before gentamicin treatment. No patients had a significant response after treatment.
In Figure 4, the SVH-V test before and after gentamicin treatment is shown. Two patients had a tilt of the SVH-V of more than 3 degrees before treatment. Seven patients changed their tilt 3 degrees or more toward the side of the lesion after the treatment and all the 4 patients changed their tilt altogether toward the side of the lesion. Overall the change was 2.2 degrees toward the side of the lesion (p = 0.010). There were no correlation between tumor size and the caloric response from side affected by the tumor before (p = 0.794, R = 0.074) or after gentamicin treatment (p = 0.335, R = −0.268). Moreover, there were no correlation between tumor size and the SVH-V value before (p = 0.794, R = 0.074) or after gentamicin treatment (p = 0.639, R = 0.127).
At the time for check-up, all patients were well compensated and displayed no spontaneous, gaze, or positional nystagmus, except for 3 patients who had a very mild nystagmus, beating toward the contralateral ear, only detectable with Frenzel’s goggles. The patients reported vertigo symptoms that culminated the week after the injections. The severity differed but no one was compelled to stay at home and be absent for work. Neither did the symptoms hinder them from doing the vestibular exercises.
Although being a small material, this is the first study on comparably normal inner ears and intratympanic gentamicin installations. The gentamicin effect on hearing levels seems to be comparable with those seen when treating ears with endolymphatic hydrops and a definite Menière’s disease (Pullens & van Benthem 2011; Gabra & Saliba 2013). Postema et al. (2008) demonstrated a decrease in PTA of 8.1 ± 18.1 dB in their material, which is similar to our results of 7.1 ± 8.5 dB decrease in PTA. However, it is possible that the hearing levels before the intratympanic injections were better in the present study, than in studies on Menière patients. In the study from Stokroos and Kingma (2004) for instance, PTA levels averaged 60 dB before treatment. Despite the possibly better hearing level, the patients in the present study did not experience greater hearing loss than in the above studies. The results are also comparable with Menière patients considered to have stage I hearing, treated with intratympanic gentamicin (Silverstein et al. 2010). The conclusion drawn from above studies has been that intratympanic gentamicin could be considered safe when treating Menière patients with serviceable hearing. Although we found the same differences in PTA levels, there was a significant hearing loss predominantly in high-frequency hearing. Thus, it would seem to benefit both researchers and clinicians if the effects on individual frequencies also were analyzed, when studying interventional effects of gentamicin on hearing as well as using multivariate statistical methods. It is also important to acknowledge the fact that the hearing levels inherently fluctuates in Menière’s disease and that the present study therefore is not directly comparable with previous reports on the potential cochleotoxicity of gentamicin. The numbers of gentamicin treatments also differ, but our results stress the importance of reducing the number of treatments, a minimum without compromising the aim of vestibular deafferentation.
The hearing results in this study were the same as those previously reported in gentamicin intervention studies performed on Menière patients. The key question is if gentamicin therapy could, thus, be considered safe to use before surgery with the intention of preserving hearing. The answer is not easy because patients with both perfect and bad hearing were affected, and the range in recorded hearing loss in association with the gentamicin intervention was very large within the present material (Table 1). Furthermore, the hearing change was within 10 dB difference in 12 of 17 patients, and 9 of 13 patients did not change their hearing classification. However, we consider it to be prudent to carefully assess the benefits of a better vestibular and postural compensation (Tjernström et al. 2009) relative to the risk for hearing loss, and exclude patients with perfect hearing (class A) from the gentamicin therapy, if the surgery aims for hearing preservation. The risk for hearing loss following gentamicin therapy should also be put in perspective to the risk for hearing loss due to the actual treatment of the schwannoma itself, be it either through microsurgery or radiosurgery. Despite the fact that hearing preservation has improved due to gentler techniques (nerve monitoring as well as refining radiation doses), any intervention still carries a great risk for hearing loss (Kari & Friedman 2012; Sarmiento et al. 2013; Yang et al. 2013; Yamakami et al. 2014).
The inner ears of schwannoma patients might not be normal. A recent review revealed cochlear pathology based on temporal bone studies (Roosli et al. 2012). They found, for example, hydrops in 6 of 22 ears, although a further 2 patients had hydrops contralaterally. It is thus conceivable that a number of patients included in the present study also had hydrops of clinical relevance.
It might be argued that the high-frequency loss seen in 2 patients could be attributed to schwannoma growth, which characteristically initially diminish high-frequency hearing. However, the hair cell destruction from gentamicin effect follows a base-apex gradient which would generate an initial high-frequency hearing loss (Forge & Schacht 2000), thus rendering the etiology indeterminable. The size of the schwannoma did not correlate to PTA level, which corroborates the finding of Lee et al. (2015). This, together with finding a similar average hearing loss as when treating Menière patients with gentamicin, strongly suggests that gentamicin caused the hearing loss in the present material. Ideally a control group should have been assessed; however, the acquisition of controls that share the same patient properties, hearing loss as well as tumor characteristics (and same tumor growth from the same initial size during the investigated time frame) as in the present study, is nigh on impossible.
The results suggest that gentamicin targets the vestibular system also in comparably normal inner ears, and result in a total uVD. However, it is a well-known fact, from treating Menière’s disease, that all patients do not become free from vertigo following gentamicin installations (Cohen-Kerem et al. 2004). This might be due to an uncertain effect or delivery of gentamicin to the inner ear (differences in middle ear mucosal status, round window thickness or adhesion, patency of Eustachian tube, or the effect of endolymphatic hydrops on ototoxicity; Berryhill & Graham 2002). The success rate might be higher in comparably normal ears since the endolymphatic hydrops as a detrimental factor is at least not as frequent (Roosli et al. 2012). Another contributing part to continuing attacks after gentamicin treatment in Menière’s disease could be that the vestibular hair cells regenerate (Forge et al. 1998), and thus the vertigo could reappear. In the vestibular PREHAB protocol however that would not be an issue, because regeneration takes time and the surgery ideally is scheduled within 2 to 3 months after the induced deafferentation.
Complications to the intratympanic treatment happened in 2 patients (persisting perforation). These were treated with conventional or fat plug myringoplasty. In both these cases, phenol (carbolic acid) was used to anesthetize the tympanic membrane which could explain the perforations (Sing 2006). Although this method has been the general practice for many years at our clinic, very few perforations have been observed.
Intratympanic installations of gentamicin, as part of the vestibular PREHAB, are well tolerated with few recorded complications and result in controlled uVD. However, there is a definite risk for loss of high-frequency hearing as a consequence of the intervention.
The authors declare no conflict of interest.
Ahmed R. M., Hannigan I. P., MacDougall H. G., et al Gentamicin
ototoxicity: A 23-year selected case series of 103 patients. Med J Aust, (2012). 196, 701–704.
Altman D. Practical Statistics for Medical Research. (1991). New York, NY: Chapman & Hall.
American Academy of Otolaryngology-Head and Neck Surgery Foundation, INC. Committee on hearing
and equilibrium guidelines for the evaluation of hearing
preservation in acoustic neuroma (vestibular schwannoma
). Otolaryngol Head Neck Surg
, (1995). 113, 179–180.
Berryhill W. E., Graham M. D. Chemical and physical labyrinthectomy for Meniere’s disease. Otolaryngol Clin North Am, (2002). 35, 675–682.
Böhmer A., Rickenmann J. The subjective visual vertical as a clinical parameter of vestibular function
in peripheral vestibular diseases. J Vestib Res, (1995). 5, 35–45.
Cohen-Kerem R., Kisilevsky V., Einarson T. R., et al Intratympanic gentamicin
for Menière’s disease: A meta-analysis. Laryngoscope, (2004). 114, 2085–2091.
Forge A., Schacht J. Aminoglycoside antibiotics. Audiol Neurootol, (2000). 5, 3–22.
Forge A., Li L., Nevill G. Hair cell recovery in the vestibular sensory epithelia of mature guinea pigs. J Comp Neurol, (1998). 397, 69–88.
Gabra N., Saliba I. The effect of intratympanic methylprednisolone and gentamicin
injection on Ménière’s disease. Otolaryngol Head Neck Surg, (2013). 148, 642–647.
Gurgel R. K., Jackler R. K., Dobie R. A., et al A new standardized format for reporting hearing
outcome in clinical trials. Otolaryngol Head Neck Surg, (2012). 147, 803–807.
Hafstrom A., Fransson P. A., Karlberg M., et al Subjective visual tilt and lateral instability after vestibular deafferentation. Acta Otolaryngol, (2006). 126, 1176–1181.
Kari E., Friedman R. A. Hearing
preservation: Microsurgery. Curr Opin Otolaryngol Head Neck Surg, (2012). 20, 358–366.
Karlberg M., Annertz M., Magnusson M. Acute vestibular neuritis visualized by 3-T magnetic resonance imaging with high-dose gadolinium. Arch Otolaryngol Head Neck Surg, (2004). 130, 229–232.
Lee S. H., Choi S. K., Lim Y. J., et al Otologic manifestations of acoustic neuroma. Acta Otolaryngol, (2015). 135, 140–146.
Magnusson M., Karlberg K., Halmagyi G. M., et al The video-impulse test enhances the possibility of detecting vestibular lesions. J Vestib Res, (2002). 11, 231.
Magnusson M., Kahlon B., Karlberg M., et al Vestibular “PREHAB”. Ann N Y Acad Sci, (2009). 1164, 257–262.
Magnusson M., Karlberg M., Tjernström F. ‘PREHAB’: Vestibular prehabilitation to ameliorate the effect of a sudden vestibular loss. NeuroRehabilitation, (2011). 29, 153–156.
Manzari L., Burgess A. M., Curthoys I. S. Dissociation between cVEMP and oVEMP responses: Different vestibular origins of each VEMP? Eur Arch Otorhinolaryngol, (2010). 267, 1487–1489.
Postema R. J., Kingma C. M., Wit H. P., et al Intratympanic gentamicin
therapy for control of vertigo in unilateral Menire’s disease: A prospective, double-blind, randomized, placebo-controlled trial. Acta Otolaryngol, (2008). 128, 876–880.
Pullens B., van Benthem P. P. Intratympanic gentamicin
for Meniere’s disease or syndrome. Cochrane Database Syst Rev, (2011). (3), CD008234.
Roosli C., Linthicum F. H. Jr, Cureoglu S., et al Dysfunction of the cochlea contributing to hearing
loss in acoustic neuromas: An underappreciated entity. Otol Neurotol, (2012). 33, 473–480.
Sarmiento J. M., Patel S., Mukherjee D., et al Improving outcomes in patients with vestibular schwannomas: Microsurgery versus radiosurgery. J Neurosurg Sci, (2013). 57, 23–44.
Schmid-Priscoveanu A., Böhmer A., Obzina H., et al Caloric and search-coil head-impulse testing in patients after vestibular neuritis. J Assoc Res Otolaryngol, (2001). 2, 72–78.
Silverstein H., Wazen J., Van Ess M. J., et al Intratympanic gentamicin
treatment of patients with Ménière’s disease with normal hearing
. Otolaryngol Head Neck Surg, (2010). 142, 570–575.
Sing T. T. Use of phenol in anaesthetizing the eardrum. Int J Otorhinolaryngol, (2006). 4.
Stokroos R., Kingma H. Selective vestibular ablation by intratympanic gentamicin
in patients with unilateral active Ménière’s disease: A prospective, double-blind, placebo-controlled, randomized clinical trial. Acta Otolaryngol, (2004). 124, 172–175.
Tjernström F., Fransson P. A., Kahlon B., et al Vestibular PREHAB and gentamicin
before schwannoma surgery may improve long-term postural function. J Neurol Neurosurg Psychiatry, (2009). 80, 1254–1260.
Tjernström F., Nyström A., Magnusson M. How to uncover the covert saccade during the head impulse test. Otol Neurotol, (2012). 33, 1583–1585.
Yamakami I., Ito S., Higuchi Y. Retrosigmoid removal of small acoustic neuroma: Curative tumor removal with preservation of function. J Neurosurg, (2014). 121, 554–563.
Yang I., Sughrue M. E., Han S. J., et al A comprehensive analysis of hearing
preservation after radiosurgery for vestibular schwannoma
: Clinical article. J Neurosurg, (2013). 119 Suppl, 851–859.
Keywords:Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
Gentamicin; Hearing; Orehabilitation; Vestibular function; Vestibular schwannoma