How serious a problem is topical aminoglycoside ototoxicity?

Although the danger of aminoglycoside ototoxicity has perhaps been overstated, it nonetheless does exist. The scope of the danger depends in part on how we define ototoxicity and whether the injury to the inner ear is partial or complete.

Cochleotoxicity

It would be fairly easy to identify cochleotoxicity if one had a pre- and postevent audiogram that clearly demonstrated an objective change in hearing. However, this is seldom the case. One might reasonably anticipate that instillation of any topical cochleotoxic agent into the middle ear would primarily affect the inner hair cells responsible for ultrahigh-frequency hearing that are closest to the round window membrane portal of entry. Ultrahigh-frequency hearing, however, is not routinely evaluated at most testing facilities. The upper limit is usually 8 kHz. Therefore, it is probable that we are missing some cases of ototoxicity at the higher frequencies.

Vestibulotoxicity

The vestibular portion of the inner ear remains to a large degree a mysterious "black box" for most clinicians. In the acute setting following a loss of unilateral vestibular function, one might see evidence of dysfunction involving vestibulo-ocular, vestibulospinal, and vestibulocerebellar reflexes. Such dysfunction from an acute loss leads to nystagmus, falling toward the side of the injury, ataxia, and other features of hemispheric cerebellar dysfunction. This is a very complicated system, and when vestibulotoxicity occurs, it is more than an all-or-none phenomenon--in other words, a spectrum of vestibular dysfunction can exist.

Recognition of clinical vestibular dysfunction, especially when chronic, can be difficult. Most otolaryngologists do not have a close familiarity with clinical tests such as the head-shake test, the oscillopsia (dynamic visual acuity) test, the high-frequency horizontal head-thrust (Halmagyi) maneuver, and the head-heave translational otolithic maneuver. These are tests that experienced neurotologists perform when they evaluate patients with suspected vestibular dysfunction. Perhaps the most robust of these tests is the Halmagyi maneuver, which specifically assesses the horizontal vestibulo-ocular reflex (VOR). The function of the VOR is to provide an equal but opposite movement of the eyes to the head so that one maintains a stable retinal image.

Conventional vestibular testing usually involves electronystagmography (ENG) and rotational chair testing. Nevertheless, problems can arise in certain circumstances. Unless an examiner uses air-caloric irrigation or a closed-loop system, he or she will be unable to assess caloric activity in an ear that has a tympanic membrane defect. Moreover, rotational chair testing addresses only low-frequency/acceleration vestibular function; it does not--and really cannot--assess the everyday high-velocity/acceleration head and corresponding eye movements we continually make.

Advances in our technical ability to measure the VOR clinically have changed the way we view this reflex. Many otolaryngologists may not be aware that we can divide vestibular function into low- and high-frequency types. This is a very important concept to remember, as caloric testing during ENG typically measures a low-frequency vestibular response and cannot measure the high-frequency vestibular responses that are more common in everyday life. Accordingly, a patient can have an isolated high-frequency vestibular loss yet still maintain low-frequency vestibular activity on caloric testing. Under these circumstances, the results of an ENG will be normal, but the patient will still exhibit clinical symptoms of vestibular dysfunction with active head movement.

When measuring high-frequency vestibular activity, we can turn to the fairly elaborate procedure first developed in primate research called scleral coil eye-movement recording. In this procedure, the patient sits in a rotational chair surrounded by a frame that produces a magnetic field around the patient. With the patient under topical anesthesia, a contact lens with a magnetic coil is inserted onto the cornea. With the chair in motion or with rapid head accelerations/impulses, the coil records the relationship of eye movement to head movement. This recording can be accomplished without the possibility of electrode slippage, which can occur during conventional electro-oculography. This is the most accurate test available today for assessing the VOR.

A loss of unilateral VOR function is relatively well compensated for as long as the other ear and the central vestibular connections are intact. Unless one is engaging in activities that significantly stress the VOR (e.g., skiing down a steep slope or playing a rapid game of squash), any visual blurring from retinal slip will be minimal. When one loses vestibular function bilaterally (a phenomenon that is usually seen with systemic aminoglycoside ototoxicity but which can also occur in patients who receive topical aminoglycoside-containing drops in the presence of a tympanic membrane perforation), compensation can never be complete, despite the best efforts of vestibular rehabilitation. Bilateral loss of high-frequency vestibular function severely affects the VOR, leading to subjective complaints of imbalance that become worse in the absence of vision and in oscillopsia (visual blurring with head movement).

What this preamble all leads up to is that our ability to detect changes in vestibular function is sometimes dependent on very advanced technology that is not available outside of a few centers in the Western world. Therefore, the recognition of a partial vestibular loss caused by an ototoxic eardrop may be difficult clinically, even in symptomatic patients.

Topical aminoglycoside ototoxicity

Several factors have an influence on whether a topical aminoglycoside will be ototoxic in a patient with a perforated eardrum:

* The tympanic membrane defect might be too small to allow passage of drops into the middle ear space, or the drops might be too viscous to penetrate the defect.

* Eustachian tube dysfunction might result in drops passing through the tube and into the nasopharynx rather than being absorbed by the middle ear.

* An accumulation of pus or mucus can act as a barrier to absorption into the inner ear, as can mucosal webs around the round window membrane.

* A highly vascular middle ear mucosa will absorb drops before they reach the round window membrane.

* The permeability of the round window membrane is certainly variable in any individual and can be affected by the health of the middle ear.

* There are differences in individual patients' susceptibility to aminoglycoside toxicity.

Perhaps the best evidence of topical aminoglycoside ototoxicity is seen when intratympanic gentamicin is administered for the treatment of incapacitating Meniere's disease. According to some meta-analyses, rates of vertigo control with this therapy range from 75 to 100%. (1-3) In these studies, hearing loss did not appear to be related to the frequency or total number of instillations or to the length of treatment. There was also some subtle evidence to suggest that treatment failure was associated with incomplete vestibular ablation rather than an attenuated vestibular response.

When our group at the University of Toronto began using topical gentamicin ablation therapy for incapacitating unilateral Meniere's disease in the 1980s, we administered a concentrated solution of 25 mg/ml three times daily for 3 days. We were able to demonstrate a significant reduction in caloric activity in well over half of all patients by comparing their pre- and post-treatment caloric values. We found that vestibulotoxicity occurred in a delayed fashion--that is, 2 to 4 days following the cessation of treatment. This suggested that there was an ongoing concentration effect in the inner ear before clinical ototoxicity manifested.

Shortly thereafter, we began to recognize that a number of patients who had used a commercially available combination gentamicin/steroid drop--3 mg/ml ofgentamicin and 1 mg/ml of betamethasone--had developed a unilateral and, in some cases, a bilateral loss of vestibular function in the presence of a tympanic membrane perforation. Since 1991, we have identified 32 patients who have experienced this inadvertent topical ototoxicity. Most of these patients had taken the gentamicin/betamethasone drops for an extended length of time (i.e., ~2 wk). Most of the patients with unilateral dysfunction were able to adapt and compensate over time. However, those with bilateral loss of function were typically incapacitated by imbalance, ataxia, and oscillopsia, and they did not improve very much.

At that point, one of my colleagues suggested that if commercially available gentamicin drops could cause ototoxicity in patients with an open middle ear space, perhaps we could use these drops to intentionally cause ototoxicity in patients with unilateral incapacitating Meniere's disease by ablating vestibular function. So we obtained audiometric/ENG data on these patients and then inserted a tympanostomy tube into the posteroinferior quadrant of the eardrum. We instructed these patients to place 5 or 6 drops of gentamicin/betamethasone drops into the ear 3 times a day until they became dizzy for 2 consecutive days and then to stop the medication. Most patients who developed intentional vestibulotoxicity during treatment of their Meniere's disease did so within 12 days, while most of those who had developed unintentional vestibulotoxicity, apparently in the treatment of middle ear disease, did so in an average of 15 days. The 3-day difference perhaps reflects the length of time it took for the inflammatory component of their middle ear disease to resolve. (4)

When we compared pre- and postcaloric responses following treatment with the low- and high-concentration gentamicin solutions, the curves were similar (figure). This implies that even at low concentrations, commercially available topical gentamicin drops can induce ototoxicity if they are administered long enough.

[FIGURE OMITTED]

Consequences of ototoxicity

Ototoxicity is not a victimless condition. Bilateral ototoxicity has a profoundly morbid effect on patients. Some are never able to return to work or even free themselves from a wheelchair. At our institution, we examined quality of life in a series of patients with systemic gentamicin-induced vestibulotoxicity. Using the Short Form 36 assessment tool, we measured various quality-of-life parameters in these patients and compared them with normal age- and sex-matched controls from among the general Canadian population. (5) These parameters included factors such as physical and social functioning, emotional and mental status, vitality, and general overall health. It was no surprise that the vestibulotoxic patients fared very poorly in comparison with the controls, particularly with regard to physical functioning.

Reporting of ototoxicity

I believe that topical aminoglycoside-induced ototoxicity overall is probably underreported. The reasons for this involve physician factors, patient factors, and genetic factors.

Physician factors. Ototoxicity goes undiagnosed, especially when a physician is evaluating only hearing or looking for evidence of cochleotoxicity only. Vestibulotoxicity will often be missed. Other recognition deficiencies include a failure to take a good history or perform adequate clinical testing. Moreover, comparative pre- and post-treatment audiometric and vestibular studies are rarely available. Finally, a physician may neglect to report ototoxicity for fear of being sued.

Patient factors. Patients may believe that their loss of vestibular function was caused by their underlying pathology rather than by its treatment. Also, many patients recover spontaneously from their loss, and the ototoxicity becomes forgotten. Similarly, some who do not recover attain adequate function through compensation and, again, any loss will be forgotten.

Genetic factors. We are learning more about the genetic basis of ototoxicity. Some people have an inherent genetic susceptibility to aminoglycoside toxicity. Certain pedigrees in Southeast Asia and the Middle East have a mitochondrial DNA mutation that predisposes them to profound nonsyndromic deafness after a short course of a systemic aminoglycoside, particularly kanamycin. Some researchers are concerned that patients who take an aminoglycoside will present later in life with a progressive sensorineural hearing loss, long after they have been exposed to the aminoglycoside. One wonders whether the same might hold true for the vestibular system, as well.

What makes all this evidence and conj ecture almost moot is that there is no need to use a topical aminoglycoside. We already have topical quinolones, which are not only safer, but also more effective.

References

(1.) Blakley BW. Update on intratympanic gentamicin for Meniere's disease. Laryngoscope 2000; 110:236-40.

(2.) Gustafson ML, Pensak ML. Inner ear perfusion therapy: An update. Current Opinion in Otolaryngology and Head and Neck Surgery 2000;8:398-402.

(3.) Cohen-Kerem R, Kisilevsky V, Einarson TR, et al. Intratympanic gentamicin for Meniere's disease: A meta-analysis. Laryngoscope 2004; 114:2085-91.

(4.) Kaplan DM, Hehar SS, Bance ML, Rutka JA. Intentional ablation of vestibular function using commercially available topical gentamicin-betamethasone eardrops in patients with Meniere's disease: Further evidence for topical eardrop ototoxicity. Laryngoscope 2002; 112:689-95.

(5.) Roland PS, ed. Antibiotic ototoxicity: Pathophysiology, management, and medicolegal implications. Ear Nose Throat J 2003;82(suppl 1):1-16.

John Rutka, MD, FRCSC