Intratympanic Gentamicin for the Treatment of Unilateral Meniere's Disease


  • Daniel M. Kaplan MD,

    1. Department of Otolaryngology, Sunnybrook and Women's College Health Science Centre and the University of Toronto, Toronto, Ontario, Canada
    Search for more papers by this author
  • Julian M. Nedzelski MD,

    Corresponding author
    1. Department of Otolaryngology, Sunnybrook and Women's College Health Science Centre and the University of Toronto, Toronto, Ontario, Canada
    • Julian M. Nedzelski, MD, Department of Otolaryngology, Suite M1.102, Sunnybrook and Women's College Science Health Centre, 2075 Bayview Avenue, Toronto ON M4N 3M5, Canada.
    Search for more papers by this author
  • Joseph M. Chen MD,

    1. Department of Otolaryngology, Sunnybrook and Women's College Health Science Centre and the University of Toronto, Toronto, Ontario, Canada
    Search for more papers by this author
  • David B. Shipp MA

    1. Department of Otolaryngology, Sunnybrook and Women's College Health Science Centre and the University of Toronto, Toronto, Ontario, Canada
    Search for more papers by this author

  • Presented at the Meeting of the Eastern Section of the American Laryngological, Rhinological and Otological Society, Inc., Pittsburgh, Pennsylvania, January 29, 2000.


Objective To determine the efficacy of intratympanic gentamicin instillation as treatment of incapacitating unilateral Meniere's disease, using a predetermined regimen with a fixed dose.

Study Design A prospective study from a single institution between 1988 and 1998.

Methods One hundred fourteen patients were enrolled in this study. Gentamicin (26.7 mg/mL) was administered three times daily for 4 consecutive days. The Committee on Hearing and Equilibrium Guidelines for Reporting Treatment Results in Meniere's Disease of the American Academy of Otolaryngology and Head and Neck Surgery (1985) were used.

Results Comprehensive data were available for 90 individuals. Complete control of vertigo was achieved in 76 (84.4%), substantial control in 8 (9.0%), limited control in 2 (2.2%), and insignificant control in 4 (4.4%) patients. Disability scores at the end of 2 years were as follows: 76 patients (84.4%) had no disability, 5 (5.6%) had mild disability, 2 (2.2%) had moderate disability, and 7(7.8%) had severe disability. Caloric testing responses, as determined using electronystagmography, were as follows: 71% of the patients had an absent ice-water response, 16% had a positive ice-water response, and in 13% there continued to be present a bithermal response. Hearing was worse in 22 patients (25.6%), unchanged in 41 (48.2%), and improved in 22 (25.6%).

Conclusions Intratympanic gentamicin administration using this particular protocol is an effective treatment option for patients with disabling unilateral Meniere's disease. Hearing loss is a distinct possibility, and patients should be advised accordingly.


The management of Meniere's disease continues to be a challenge. Patients present with a wide spectrum of symptoms, of which vertigo is by far the most distressing. Although much has been written about the pathophysiology of Meniere's disease, there is currently no curative treatment and therapy is primarily aimed at the relief of vertigo. The majority of individuals respond to medical treatment consisting of salt restriction, diuretics, vasodilators, and symptomatic therapy for the nausea and diaphoresis associated with spells of vertigo. Definitive treatment is surgical and has consisted of either labyrinthectomy or selective vestibular neurectomy, depending on hearing acuity and other factors.

Since the late 1970s the selective ototoxic effects of aminoglycosides have been successfully exploited as treatment for these patients. Of these, gentamicin and streptomycin have been shown to be more damaging to the vestibular hair cells (vestibulo-ototoxic) than to cochlear hair cells (cochleo-ototoxic). 1,2 In addition, a number of animal studies have suggested that aminoglycosides cause ototoxic damage to the dark cells within the stria vascularis, which is thought to produce endolymph, therefore theoretically reducing endolymph volume and providing additional benefit. 3,4

Transtympanic instillation of aminoglycosides allows for the isolated treatment of one ear, without the potential of systemic effect. The drug accesses the inner ear primarily through the round window membrane with additional amounts absorbed through the annular ligament of the oval window, via blood or lymph vessels or through the tiny lacunae in the bony wall. 5

Schuknecht 1 was the first author to report on the instillation of aminoglycosides into the middle ear in 1957. He treated eight patients with streptomycin. Of these, five patients realized vertigo control; however, all were deafened. Schuknecht's experience was mirrored by Silverstein 6; since then, streptomycin has not been used in this form. Since the late 1970s, intratympanic gentamicin has been used by many investigators worldwide, employing a variety of protocols. 5,7–13

The intent of this study was twofold: to examine the efficacy of our protocol in controlling the disabling vertigo associated with unilateral Meniere's disease and to highlight the effects on hearing. Results are reported using the using Committee on Hearing and Equilibrium Guidelines for Reporting Treatment Results in Meniere's Disease of the American Academy of Otolaryngology—Head and Neck Surgery (AAO-HNS). 14


All patients enrolled in this study had a diagnosis of unilateral “definite Meniere's disease” as defined by the 1985 Committee on Hearing and Equilibrium Guidelines for Reporting Treatment Results in Meniere's Disease. 14 Each patient had been treated previously with the following: a low-salt diet, diuretics, and symptomatic therapy over a minimum period of 6 months. Only after failing this treatment were they considered for intratympanic instillation of gentamicin. Other therapeutic considerations (i.e., vestibular nerve section and labyrinthectomy) were discussed with each patient. All patients underwent a detailed audiological assessment including a pure-tone audiogram, speech reception threshold testing, and speech discrimination tests. Electronystagmography (ENG) was performed, including bithermal/ice-water caloric tests.

The method of application of the intratympanic gentamicin and follow-up protocol have been described in detail 7,15,16 and are mentioned only briefly. Gentamicin (40 mg/mL) was buffered to a final concentration of 26.7 mg/mL (pH 6.4). After appropriate local infiltration of the external auditory canal with lidocaine, a myringotomy in the posteroinferior quadrant was carried out. A “T” type of ventilation tube attached to a butterfly type of catheter tubing was inserted into the tympanic membrane, and 0.7 to 0.8 mL of the gentamicin solution delivered at each dose. The drug was instilled through the catheter three times daily, the first dose delivered at noon in our clinic and the next two doses at 7:00 pm and 7:00 am by a family member or friend. Administration to the ear was carried out with the patient supine, the affected ear upright. The patient was instructed not to swallow (tissues were supplied for saliva) and to remain supine for 30 minutes. Patients were treated for 4 consecutive days (total dose of approximately 208 mg).

Patients were seen daily. At each visit, symptoms were recorded and each patient was examined for spontaneous and gaze-associated nystagmus. Tandem gait was assessed. Bone conduction audiometry was performed and, if no difficulties were identified, the noon dose of gentamicin given. Treatment was discontinued if there was nystagmus, deterioration of tandem gait, or worsening of hearing by more than 10 dB on three consecutive frequencies. On the fifth day the ventilation tube was removed and an audiogram performed. The patient was followed at 1-, 3-, 6-, 12-, 18-, and 24-month intervals for the first 2 years, and yearly thereafter. On each follow-up visit, in addition to an updated history, neurotological examination was performed, as well as audiometry and ENG recording of the caloric response. The latter always included a calculation of the excitability difference (ED). When responses were less than 5° per second, using the conventional bithermal stimuli, ice-water stimulation was employed. Responses using an ice-water stimulus were recorded as being present (positive) or absent (negative).

For patients to be included in the study, a minimum 24-month follow-up was required. The 1985 Committee on Hearing and Equilibrium Guidelines for Reporting Treatment Results in Meniere's disease 14 were used, rather than the 1995 guidelines, because many of the patients were treated before 1995. Vertigo control was calculated by obtaining a numerical value according to the following formula: average number of spells per month during the 24-month posttreatment period divided by the average number of spells per month during the 6 months before treatment and multiplied by 100. Depending on the numerical value, vertigo control was judged to be complete (0), substantial (1–40), limited (41–80), insignificant (81–120), or worse (>120). The disability score was determined, together with the patient, using the most appropriate sentence describing his or her disability and was defined as none, mild, moderate, or severe.

Caloric test responses were reported before and after treatment (at 1 mo, 1 y, and 2 y). These were defined as follows: 1, negative ice-water test response; 2, positive ice-water test response only; 3, ED of 75% to 100%; 4, ED of 50% to 74%; 5, ED of 25% to 49%; and 6, ED of less than 24%.

Pure-tone hearing acuity was calculated as the average thresholds at 500, 1000, 2000 and 3000 Hz (pure-tone average [PTA]). Hearing was considered worsened or improved when a PTA of more than a 10 dB was noted, or a 15% change in speech discrimination. Because some of the patients had nonmeasurable hearing, comparison between average pretreatment and posttreatment hearing was not used. Hearing was also reported, using the staging system devised by the Committee on Hearing and Equilibrium for the Diagnosis and Evaluation of Therapy in Meniere's Disease of the AAO- HNS in 1995. 17 Stages were defined by PTA only, as follows: stage 1, 25 dB or less; stage 2, 26 to 40 dB; stage 3, 41 to 70 dB; and stage 4, greater than 70 dB.

Data were imported from an Excel (Microsoft Office 1997) spreadsheet to the GraphPad Instat program (GraphPad Software, Inc., San Diego, CA) for statistical analysis of demographics, vertigo control, disability, hearing change, and caloric test response. The χ2 test and Fisher's exact test were used to compare frequencies. The confidence interval (CI) was 0.05.


Between June 1988 and December 1999 a total of 121 patients with unilateral Meniere's disease were treated at the Sunnybrook and Women's College Health Science Centre of the University of Toronto (Ontario, Canada). Seven patients were excluded from the present study because they were treated with once-weekly injections. The remaining 114 patients were treated with a fixed-dose protocol of 12 injections over a 4-day period. Five of these patients had been treated within the preceding 2 years, 18 patients were lost to follow-up, and 1 patient had treatment stopped after administration of the ninth dose. A total of 90 patients who had undergone the fixed-dose protocol and a minimum 2-year follow-up constituted the study group. Included in this group were five treatment failures in patients who did not have a 24-month follow-up because they underwent transmastoid labyrinthectomy/vestibular neurectomy.

A comparison of the cohort of 19 patients lost to follow-up (including the patient who stopped the treatment) with those of the group under study with respect to age, gender, pretreatment hearing, and vestibular function did not reveal any significant differences (P values were P = .83, P = .34, P = .16, and P = .29, respectively).

An overview of the data for each of the 90 patients is presented in Table I. Individuals are numbered chronologically according to the date of the first treatment. Demographic data are presented in Table II. The only complication to date has been a single persistent tympanic membrane perforation.

Table Table 1.. Patient-Specific Data.
original image
Table Table 1.. Continued.
original image
Table Table 2.. Patient Demographics.
original image

Vertigo Control and Disability Scores

Seventy-six patients realized complete control of vertigo, eight had substantial control, two had limited control, and four had insignificant control (Table III). Seventy-six patients volunteered having no disability; five, mild disability; two, moderate disability; and seven, severe disability (Table IV). Table V highlights the relationship between vertigo control, which is perhaps a more objective measure, and disability. Seventy-four of 76 patients with complete control of vertigo reported no disability. Of interest, and somewhat of a paradox, is the finding that seven patients viewed their disability as severe despite vertigo control spanning three categories.

Table Table 3.. Vertigo Control.
original image

AAO-HNS Committee on Hearing and Equilibrium Guidelines for Reporting Treatment Results in Meniere's Disease (1985). 14

*Numerical value = Average # of definitive spells per mo. (24 mo after therapy)/Average # of definitive spells per mo. (6 mo before therapy) × 100.

Table Table 4.. Disability Score.
original image

AAO-HNS Committee on Hearing and Equilibrium Guidelines for Reporting Treatment Results in Meniere's Disease (1985). 14

Table Table 5.. Relation Between Vertigo Control and Disability Score (Two Years After Treatment).
original image

Vertigo control was not associated with age (P = .09), sex (P = .12), duration of symptoms (P = .12), type of previous treatment (medical treatment only, gentamicin or surgery) (P = .13), or the number of intratympanic treatments (P = .12).

Fourteen patients were retreated once, and three patients were retreated twice. On average, these individuals were retreated 7 months (range, 3–20 mo) after their initial treatment. Three of the 17 patients did not achieve complete or substantial vertigo control, and one of these underwent a labyrinthectomy. This subgroup of patients achieved a vertigo control rate of 82%, which was statistically similar to the vertigo control rates of patients who had been treated once. None of the retreated patients had an hearing loss.

Of the five patients who underwent surgical treatment, four patients had had stage 4 hearing loss (PTA >70 dB) in the affected ear and underwent transmastoid labyrinthectomy and the other patient had a vestibular nerve section. Two of these patients (patients 4 and 30) had substantial vertigo control.

Seventy-one percent of the patient population had an absent ice-water test response, 16% had a positive ice-water test response, and in 13% a bithermal response continued to be present.


Pretreatment and 2-year posttreatment audiograms were available for all patients apart from the five who underwent surgery. Audiograms performed 1 month after treatment revealed the following: hearing was unchanged in 44 patients (51.2%), improved in 13 (15.1%), and worsened in 29 (32.5%). At 2 years, hearing remained unchanged in 41 patients (48.2%), improved in 22 (25.9%), and worsened in 22 (25.9%) (Fig. 1). Of the 22 patients with worsened hearing, 18 were observed to have this finding at the 1-month posttreatment interval. Based on the staging criteria most recently suggested by the 1995 Committee on Hearing and Equilibrium for the Diagnosis and Evaluation of Therapy in Meniere's Disease, 17 hearing was unchanged in 43 patients (50.6), improved in 21 (24.6%), and worsened in 22 (25.8%) (Fig. 2). Table VI highlights changes in the individual stages before and after treatment.

Figure Fig. 1..

Hearing acuity 1 month and 24 months after treatment according to the Committee on Hearing and Equilibrium Guidelines for Reporting Treatment Results in Meniere's Disease. 14 Vertical stripes = worse; solid grey = same; horizontal stripes = better.

Figure Fig. 2..

Hearing status by stages before treatment and 1 and 24 months after treatment according to the Committee on Hearing and Equilibrium Guidelines for Reporting Treatment Results in Meniere's Disease. 14 Dotted pattern = stage 1 (pure-tone average [PTA] < 21 dB); horizontal stripes = stage 2 (PTA 21–40 dB); vertical stripes = stage 3 (PTA 41–70 db); grey = stage 4 (PTA > 70 dB).

Table Table 6.. Relation Between Pretreatment and Posttreatment Hearing According to Hearing Stage. 17
original image

Summary: Worse: 22/85 = 25.8%; same: 43/85 = 50.6%; better: 21/85 = 24.6%.

Rows represent pretreatment hearing and columns represent hearing 2 years after treatment. The shaded areas illustrate the number of patients whose hearing was unchanged. Cells below boxed areas illustrate the number of patients with improved hearing and cells above signify the number of patients with worsened hearing.

Among the subgroup of patients who had been retreated with intratympanic gentamicin, there were no patients who realized worsened hearing. The stages of pretreatment hearing in the 22 patients who experienced worsened hearing were as follows: stage 1 in 1 patient, stage 2 in 3 patients, stage 3 in 13, and stage 4 in 6. This distribution pattern was similar to that of patients who did not realize worsening of hearing.

The caloric test responses of patients with hearing loss at the end of 2 years were as follows: 19 of 22 had negative ice-water test responses and 3 of 22 had positive ice-water test responses only. None had a residual bithermal response. These results demonstrate significantly reduced vestibular function as elicited by the caloric test responses in these individuals, compared with patients who did not realize hearing loss (P = .0403).

A total of 14 patients were deafened (16.4%) (no response at >95 dB). These patients had the following initial hearing acuities: stage 1 in one patient, stage 2 in two patients, stage 3 in seven, and stage 4 in four patients.


The intent of intratympanic instillation of gentamicin as treatment for refractory unilateral Meniere's disease is to stop or reduce the frequency and severity of vertigo attacks by altering the pathophysiology of the vestibular end organ, while attempting to preserve hearing. Successful treatment may still be associated with head movement–induced lightheadedness, as a consequence of the resultant unilateral vestibular loss and incomplete central compensation. Patients must be forewarned of this possibility.

A wide variety of administration methods exist. These differ in delivery technique, dose, frequency, and end point. Gentamicin has been delivered into the middle ear in the following ways: through a ventilation tube, by means of a small-diameter catheter, or by repeated transtympanic injections. Choosing a delivery system is usually based on the frequency of instillation. A transtympanic catheter is more convenient for the patient and the physician when delivering therapy several times daily, whereas transtympanic injection is more appropriate for patients receiving once-weekly treatment.

In the main, two treatment protocols for intratympanic instillation of gentamicin have evolved. These consist of the titration or variable-end-point technique and the fixed-dose or predetermined-dose regimens. The intent of both is to selectively destroy vestibular hair cells while minimizing cochlear hair cell damage. Schuknecht, 1 in his original description, used intratympanic injections of streptomycin to ablate vestibular function. The end point of this treatment paradigm was determined by the onset of nystagmus; however, profound hearing loss was a feature in five of eight of these patients. In subsequent studies, in which gentamicin had replaced streptomycin, the end point of treatment was redefined. Symptoms and signs associated with inner ear dysfunction (i.e., disequilibrium, motion intolerance, ataxia, and nystagmus) have become the clinical criteria for determining the end point. 8–10,18 In clinical practice, monitoring these symptoms and signs with the express intent of determining the optimal termination point so as to induce maximal vestibular suppression without hearing loss is extremely difficult. For example, the symptom of “continuous vertigo” during once-weekly injections has been considered by one author as a drug effect, necessitating the termination of therapy, 19 and by another as evidence of poor control, requiring ongoing treatment. 20

The intent of using a fixed protocol is to achieve a similar clinical end point within a predetermined time interval. Although the protocol is fixed, treatment is immediately terminated if ototoxic symptoms appear before completion of the treatment regimen.

The distinction between titration and fixed protocols may not be entirely clear, because some titration protocols deliver a minimal number of injections and then repeat them as necessary. Similarly, a fixed protocol can be repeated. Both have been employed using different time intervals between injections, although in fixed protocols therapy is administered at least once daily whereas titration protocols are more commonly delivered on a weekly basis. The protocol used within our clinic has, over time, proven to be well tolerated, easy to administer, and convenient, both to the patient and the physician. Apart from the first 20 patients, all the rest have been treated in an outpatient setting. Ototoxic symptoms rarely occur during the first 4 days of drug instillation and generally begin 3 to 7 days after completion of treatment. Our observation has been echoed by other authors using a similar protocol. 11

Table VII highlights a comparison of the findings in the present study with those of several recently published studies, each with a minimum of 25 patients enrolled. Our study reveals that 93% of patients realize complete and substantial control of vertigo, and as such, it compares favorably with others. Five patients with a follow-up of less than 2 years are included because all underwent surgical treatment as a result of failure to control vertigo. Current reporting does not accommodate individuals who have undergone definitive surgical treatment: labyrinthectomy or vestibular nerve section within 2 years of treatment. Exclusion of such individuals clearly holds the potential for bias. Other authors have similarly highlighted this problem. 18

Table Table 7.. Summary of Protocols and Results of Intratympanic Gentamicin Therapy for Unilateral Meniere's Disease.
original image

Only recent studies with a minimum of 25 patients that use AAO-HNS guidelines are presented.

*The single dose is determined by the final concentration after buffering multiplied by the volume injected into the tympanic membrane and had to be estimated in some of the reports.

Of interest, two of the five patients who underwent surgical intervention had substantial vertigo control but, by virtue of occupation (heavy equipment operator [patient 4]) and continued drop attacks (patient 30), had zero tolerance for symptoms and underwent surgery.

Kaasinen et al. 21 used a titration protocol and reported that 44 of 93 patients had to be retreated, but no mention was made of their results. Current AAO-HNS reporting guidelines for Meniere's disease do not address the issue of retreated patients. Regardless of the type of protocol that is used, difficulty exists in determining what constitutes a definitive end point and, consequently, makes the issue of retreatment a difficult one to address. For example, when using a titration protocol, considerable latitude exists concerning when treatment is considered to be finished. A “topping up” for gentamicin several weeks or longer after the last dose has been given, when using a titration protocol, is commonly performed. This is in contradistinction to a fixed protocol in which treatment is defined over a matter of days. We are of the opinion that if a patient has not realized vertigo control within 6 months of initial treatment, retreatment is a reasonable recommendation.

Hearing loss has been reported to occur in 10% to 47% of cases of Meniere's disease in various publications. 9,12,13,18–21 Twenty-six percent of our patients had worsened hearing. It has been our experience that hearing loss tends to make itself apparent in the first posttreatment audiogram at the 1-month interval (18/22 patients). Five patients who underwent surgery during the first 2 years of follow-up were not included in our report of hearing results. Because none of them had worsened hearing in the interval between treatment and surgery (4–12 mo) and patients who developed hearing loss usually did so in the first month after treatment, it is not likely that their hearing would have changed significantly. The overall incidence of hearing loss in our patients was comparable to other published series; however, profound hearing loss was higher.


Intratympanic gentamicin therapy based on a 4-day, fixed-dose protocol in an outpatient setting reliably achieves complete or substantial vertigo control in 93% of individuals with intractable unilateral Meniere's disease. The advantages of a fixed-dose protocol are primarily those of convenience and time management for both the patient and treating physician. Hearing loss is a distinct possibility, irrespective of the treatment protocol, when using intratympanic gentamicin, and patients must be advised accordingly. With this protocol, retreatment is a distinct option, because hearing loss has not been observed as a consequence.


The authors acknowledge the following individuals who, in their capacity as clinical neurotology fellows, ensured the continuity of this ongoing study: A. G. Pfleiderer, D. A. Schessel, G. E. Bryce, C. M. Chiong, J. Anderson, D. J. Commins, S. W. Hone, M. Al- Zuheid, and A. A. Al-Abidi.