Topical aminoglycosides? No. The case against using these agents in chronic ear disease. (Debate).

Dr. Haynes and I will debate the following resolution:

Be it resolved that the administration of topical aminoglycoside-containing drugs into the middle ear should be avoided entirely.

I am not usually endeared to extreme points of view, but I am going to argue that it is extremely unwise to use aminoglycoside-containing otic drops to treat middle ear disease, especially considering what we know about inner ear toxicity--not to mention what lawyers know about it.

There is no question that topical aminoglycosides have historically had a role to play in the treatment of external otitis, post-tympanostomy tube otorrhea secondary to an infection or tubal granuloma, and so-called safe chronic suppurative otitis media, with or without cholesteatoma. And topical aminoglycosides are certainly effective against the usual pathogens: Pseudomonas aeruginosa, Staphylococcus aureus, Escherichia coli, Proteus species, and anaerobes.

Nevertheless, all aminoglycosides have a dark side: They can cause ototoxicity. In the case of systemic gentamicin, ototoxicity appears to be primarily related to the duration of treatment, especially when the treatment course exceeds 10 to 14 days. It is also important to realize that gentamicin-induced ototoxicity tends to be primarily vestibular (figures 1 and 2), although cochleotoxicity is seen as well.

Mechanism of aminoglycoside transport

Not all aminoglycosides are vestibulotoxic. Streptomycin and gentamicin, for example, are primarily vestibulotoxic, while neomycin appears to be primarily cochleotoxic. The difference in the types of toxicity is related to the particular aminoglycoside molecule's end attachment.

We have good scientific evidence that when a topical aminoglycoside reaches the middle ear, it can pass into the inner ear, primarily through the round window membrane. This is not merely a passive process of diffusion; there is an active transport mechanism that delivers the aminoglycoside across the membrane itself. The existence of active transport has been confirmed in numerous animal studies, as observed in the passage of tagged substances, including horseradish peroxidase, (131) I-labeled ions, albumin, and neomycin. The existence of this mechanism has also been confirmed by documented observations of histopathologic changes in the inner ear (e.g., loss of hair cells and degeneration of the crista ampullaris of the semicircular canals) and by the observation of in vivo electrophysiologic changes during auditory brainstem-evoked response testing, electrocochleography, and otoacoustic reflex testing.

Aminoglycoside ototoxicity in humans

Despite our knowledge that aminoglycosides are toxic to the inner ears of animals, evidence of topical ototoxicity in humans has been more difficult to discern. Nevertheless, this does not mean that aminoglycoside-induced ototoxicity in humans does not occur. In fact, I suspect that it is probably more common than most physicians believe. For various reasons, topical ototoxicity is probably under-reported by both patients and physicians.

In order to appreciate why topical aminoglycoside-induced ototoxicity seems to be so uncommon, it is useful to look at certain middle ear anatomic factors, as well as individual patient and physician factors.

Anatomic factors. Aminoglycoside entry into the middle ear is influenced by the size of the tympanic membrane perforation and the viscosity of the preparation itself. Drops cannot easily pass through perforations that are very small. Likewise, drops that are very viscous cannot pass through most perforations, regardless of their size.

The status of the eustachian tube also has an effect on inner ear absorption. In a patient whose eustachian tube is open, drops that reach the middle ear space will not reside there long enough to be transported to the inner ear; instead, they will migrate down the eustachian tube.

Other deterrents to absorption into the inner ear are physical barriers, such as pus and mucus, which can protect the round window membrane. A significant number of patients have mucosal webs in the round window niche that protect the round window membrane and prevent drops from being absorbed by the inner ear. Finally, variability in the permeability of an individual's round window membrane has an effect on which patients might absorb drops into the inner ear.

Patient factors. One reason topical ototoxicity is probably under-reported is that patients appear to be resigned to the fact that any hearing or vestibular loss they might experience is the result of their underlying pathology rather than its treatment. Second, in a somewhat analogous fashion, recovery of vestibular function occurs in up to 50% of patients who discontinue systemic gentamicin therapy as soon as symptoms of ototoxicity occur; with their pathology resolved, these patients often pay little attention to the ototoxicity. Third, patients who experience a vestibular loss on only one side are usually able to compensate for it if their contralateral side is functioning well; again, if the problem is manageable, the ototoxicity is not as noticeable. Finally, some patients are simply more susceptible to ototoxicity than are others.

Physician factors. We physicians have not done a good job of recognizing topical ototoxicity, primarily because we fail to entertain the possibility. Also, we do not take a thorough enough history, and we do not conduct the bedside tests of clinical vestibular function that would allow us to draw a conclusion as to whether a vestibular loss is present. These tests--specifically, the high-frequency head-thrust (Halmagyi) maneuver, the head-shake test for nystagmus, and the oscillopsia test--can be easily performed in the clinical setting. Finally, we rarely have a pre-event audiogram or electronystagmogram that serves as a baseline, and most of our offices are not equipped for air caloric, closed-loop caloric, or rotational-chair testing.

Investigational gentamicin ototoxicity

We at the University of Toronto became aware that topical gentamicin was primarily toxic to the vestibular portion of the inner ear in the late 1980s, when we began using it to treat patients who had incapacitating unilateral Meniere's disease. Since then, our method for chemical ablation of vestibular function with intratympanic gentamicin has become recognized as an accepted form of treatment for Meniere's disease. (1) Reviews by Blakley (2) and Gustafson and Pensak (3) recently documented that intratympanic gentamicin controlled vertigo in 80 to 100% of patients with Meniere's disease. They found that some degree of cochleotoxicity did occur, but for the most part it did not appear to be related to the frequency or length of treatment or to the amount of the overall dosage. Some patients also experienced a reduction in tinnitus and aural pressure. Of interest, we now seem to have some evidence that the incidence of failed vertigo control is higher in Meniere's disease patients whose vestibular suppression is incomplete.

During our initial gentamicin ablations in the late 1980s, we delivered approximately 1.5 ml of a buffered gentamicin solution that contained approximately 27 mg/ml of gentamicin. (This concentration is approximately 10 times stronger than the gentamicin in commercially available drops in the United States.) We passed the solution through a simple butterfly catheter that had the needle end cut off and then through polyethylene tubing that had been placed through a myringotomy incision. We administered three instillations per day for 3 days, and then we removed the catheter. The gentamicin would pool in the round window niche and would then be absorbed in the inner ear. An important point to keep in mind is that although our patients rarely became dizzy while these ablations were being performed, many became severely dizzy 1 to 3 days later in a delayed fashion.

We were able to objectively determine that the gentamicin was effective in relieving the symptoms of Meniere's s disease. Although Meniere's s disease is a fairly capricious condition that is marked by periods of remission and exacerbation, we were able to document that most of the 24 patients we tested experienced statistically significant deafferentation of vestibular function, as measured by changes in caloric excitability differences (figure 3). Well over half of these patients exhibited no response to ice-water calorics.

Inadvertent gentamicin ototoxicity

What drew our attention to the fact that gentamicin drops can be toxic if they are used long enough is the fact that three consecutive patients came to our institution complaining of ataxia, imbalance, and oscillopsia as a result of a bilateral peripheral vestibular dysfunction following topical gentamicin therapy. All three patients had had defects in their tympanic membranes, and all had used a combination gentamicin/betamethasone product that is available in Canada and is the most commonly prescribed ototopical agent in Europe. This product contains 3 mg/ml of gentamicin and 1 mg/ml of betamethasone. All three patients had used these drops for a prolonged period, even after their ears had stopped discharging. Fortunately, pre-event audiometry results were available on all three patients. Follow-up audiometry revealed that there was no significant change in sensorineural reserve.

It is difficult to think of many factors other than gentamicin toxicity that could have caused these patients' bilateral peripheral vestibular loss, but we nevertheless carefully ruled out other causes by performing extensive immunologic and serologic testing, intracranial imaging, advanced brainstem-evoked response audiometry, and vestibular testing. The only conclusion we could reach was that the topical gentamicin had caused the ototoxicity. At that point, everything we thought we knew about topical gentamicin and its effects on the inner ear seemed to fall into place.

Since then, we have now identified 30 patients whom we believe have been affected by inadvertent ototoxicity caused by commercially available gentamicin/betamethasone drops. Most of these patients had used these drops for approximately 14 days on average. For the most part, their toxicity appeared to be vestibular rather than cochlear, and I suspect that this is the main reason we failed to recognize it clinically. When sensorineural hearing loss was present, it usually was identified in the higher frequencies.

When we first published our findings, they were quite controversial. (4) We heard criticisms that perhaps the vestibular loss was the result of the endotoxins or exotoxins from the infection itself or the result of a coincidental vestibular neuronitis. Others speculated that the vestibular loss had been present prior to gentamicin therapy but had gone unrecognized. All of these concerns were valid.

We faced a dilemma as to how we could prove our theory that commercially available gentamicin preparations are ototoxic. From an ethical standpoint, of course, a randomized, controlled, double-blind, prospective trial with normal volunteers was out of the question. Then one day, one of my colleagues suggested that if commercially available gentamicin could cause ototoxicity in patients who had tympanic membrane defects, then perhaps it could ablate vestibular function in patients who had incapacitating Meniere's disease, and thereby prove our hypothesis. And that is exactly what happened.

Intentional gentamicin ototoxicity

For our study, (5) we enrolled 20 patients with classic unilateral Meniere's disease as defined in criteria established by the American Academy of Otolaryngology--Head and Neck Surgery. We obtained baseline audiograms and electronystagmograms. To simulate the conditions under which our patients had experienced inadvertent ototoxicity, our patients with Meniere's disease had ventilation tubes inserted that were placed into the posteroinferior quadrant of the tympanic membrane in the affected ear. These patients were then instructed to instill four drops of commercially available gentamicin/betamethasone four times a day, a regimen that is similar to the one used for the treatment of a discharging ear. They were asked to continue taking the drops until they became dizzy for 2 consecutive days, and then to stop.

One to 6 months following ablation, we performed repeat audiography and electronystagmography. What we found fascinated us (figure 4). Most patients who became dizzy did so approximately 12 days after they had begun therapy; the earliest episode of dizziness occurred on day 7. Of the 20 patients, 15 experienced a significant change in caloric activity, 10 did not respond to ice-water calorics, 10 experienced a significant worsening of their hearing, and eight continued to have a small perforation; only one patient showed any improvement in hearing.

This study proved to us that topical antibiotics could pass through a ventilation tube, into the middle ear, then into the inner ear. The fact that toxicity occurred in a delayed fashion suggested that the gentamicin in the inner ear had concentrated gradually. Of course, we were disappointed that half our patients experienced a deterioration of their hearing. We suspect the explanation for this is that ears in patients with Meniere's disease are probably more sensitive to the effects of gentamicin than are normal ears.

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

References

(1.) Nedzelski JM, Schessel DA, Bryce GE, Pfeiderer AG. Chemical tabyrinthectomy: Local application of gentamicin for the treatment of unilateral Meniere's disease. Am J Otol 1992; 13:18-22.

(2.) Blakley BW. Update on intratympanic gentamicin for Meniere's disease. Laryngoscope 2000;l10:236-40.

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

(4.) Marais J, Rutka JA. Ototoxicity of topical eardrops. Clin Otolaryngol 1998;23:360-7.

(5.) 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.