NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Research Council (US) Committee on Hearing, Bioacoustics, and Biomechanics. Tinnitus: Facts, Theories, and Treatments. Washington (DC): National Academies Press (US); 1982.

Cover of Tinnitus

Tinnitus: Facts, Theories, and Treatments.

Show details


Over the years the treatments attempted for tinnitus have covered the range from vitamins and drugs to hypnosis and masking, but far and away the most common treatment—and in terms of absolute numbers helped, probably the most effective one—has been simple reassurance from a hearing professional that mild tinnitus is not a rare disorder, not necessarily a harbinger of imminent deafness, not a symptom of serious brain disorder, etc. One specialist estimates that explaining the problem is a sufficient solution for 95 percent of the tinnitus sufferers he sees (CIBA Foundation, 1981:273), and this statistic presumably ignores those patients satisfied by the explanations of general practitioners or others prior to referral to specialists.

When they are being precise, medical specialists often reserve the use of the word treatment for actions taken to actually cure a disorder or disease, as opposed to actions taken only to relieve the pain or distress of the disorder or disease. The latter is technically termed palliation or palliative treatment. As will be seen, in these terms essentially every procedure described below should be viewed as (at most) a palliative, even though the word treatment may be used in discussing it.

Psychological Intervention

As is true for many other serious ailments, tinnitus has the potential to create or to intensify psychological problems. Indeed, it might be argued that, in the past at least, tinnitus was more likely than many other maladies to have psychological concomitants because of the unresponsive and apparently uncaring reaction of medical professionals to the problem. A patient diagnosed as having cancer, for example, may have a strong psychological response to the news, but typically he soon begins to receive medical treatments of various sorts for the problem. By contrast, all too often the tinnitus sufferer was offered no treatment of substance and no hope of one in the future. It is fully believable that after one or two experiences of this sort, the frustrated tinnitus sufferer would begin to evidence behaviors in accord with psychological interpretations of the origin of the tinnitus. That is, it is imperative to realize that the plight of tinnitus sufferers has, until very recently, been largely ignored, and, as a consequence, the discovery of variations from "normal" in their psychological profiles cannot be unequivocally identified as cause or effect of their tinnitus problem.

Considerable evidence and simple logic indicate that tinnitus is a symptom of various physiological anomalies of the auditory and other systems and that these conditions probably strike people of all personality types indiscriminately. This is not to say that, once present, a tinnitus will not be better handled by one person than another, nor that transient psychological difficulties cannot cause people to focus on or exaggerate a preexisting or new physiological malady; the point is that psychological makeup is probably a minor contributing factor to the underlying anomaly. Ambrosino (1981) agrees that there is no one personality type more likely to contract severe tinnitus, but that once contracted, tinnitus can affect personality. House (1981) studied the personalities of 150 seriously afflicted tinnitus patients, and Hazell (1981a) administered personality inventories to his 200 patients, but no attempts were made to partition cause from effect.

The possibility of interplay between psychological and physiological factors is obvious when stress is considered. (Many tinnitus sufferers report exacerbation of their problem during times of stress.) It is known that psychological stress in a person's private or professional life can produce concomitant physiological changes. If some of these changes are responsible for a new or an enhanced tinnitus, it is understandable that a person might come to focus on this obvious symptom as the source of all his or her problems. Untangling this web might be complicated indeed, and both too time-consuming for and beyond the training of the typical hearing professional. The temptation is often great to refer such cases to psychotherapists; in the past this course has often been pursued too hastily—and ineffectively—since many patients balk at such treatment for what they strongly regard to be a physiological problem. A treatment procedure aimed at general relaxation—for example, biofeedback or just instructions to simply "slow down"—has appeal in such cases. Anecdotes abound about tinnitus symptoms diminishing or disappearing upon the natural disappearance of stress-inducing factors in a patient's life.

Surgery for Tinnitus

Tinnitus is known to be a frequent, early symptom of tumors of the internal auditory meatus and the cerebellopontine angle and of a glomus jugulare. It also can herald the onset of otosclerosis or indicate the presence of vascular anomalies of various sorts. Clearly, several of these conditions require surgical attention, and in the process, the tinnitus might be alleviated. But such surgery should be distinguished from that performed solely as a treatment for a severe tinnitus.

Over the years, various surgical techniques have been attempted as a last resort to alleviate debilitating tinnitus. These include tonsillectomy, sectioning of the vestibulo-facial anastomoses (Fisch, 1970), excision of the tympanic plexus (tympanosympathectomy; Lempert, 1946), excision of the main trunk of the vestibular nerve (Fisch, 1970), sectioning of the cochlear nerve (Fisch, 1970), and dorsal sympathectomy (Johnson, 1954), to name a few. Results have been unpredictable at best, and success rates have been low.

House and Brackmann (1981) reviewed reports of the effects on tinnitus of various surgical procedures performed for tinnitus relief, as well as for reasons other than tinnitus relief. Following successful removal of an acoustic neuroma, about 40 percent of the patients felt that the tinnitus was better and about 50 percent felt it was worse. Following successful stapedectomy, 74 percent reported the tinnitus to be absent or reduced. Following translabyrinthine section of the eighth nerve, about 45 percent of the patients felt that the tinnitus was absent or improved.

As long ago as 1928, Jones and Knudsen were advising against sectioning of the eighth nerve and destruction of the cochlea as treatments for intractable tinnitus—not because of a concern for the loss of useful hearing, but because of the general ineffectiveness of these procedures. Sectioning of the eighth nerve and labyrinthectomy are described as partly or wholly effective against tinnitus in about 50 percent of the cases (Hazell, 1979c), but it must be recalled that these percentages are presumably based upon small numbers of seriously debilitated patients. Thus, failure must be particularly disappointing to the patients. We have no way of estimating how many operations of these types are still performed, but they continue to appear to be ill advised solely as treatments for tinnitus. As Hazell (CIBA Foundation, 1981:213) has pointed out, cutting the eighth nerve may obviate the later use of other, possibly more effective treatments yet to be developed. (Surgical procedures advocated for Meniere's Disease and its attendant tinnitus were discussed in the section ''Meniere's Disease" in Chapter 2.)

Exposure to Intense Sound

From personal experience, it appears that any brainstorming session on tinnitus that includes people of varying levels of knowledge about the topic is likely to eventually produce a question of the following sort: Might it be possible to eliminate a narrowband tinnitus by selectively destroying the spectral region giving rise to it? The question often takes the more explicit form: Might exposure to a narrowband or tonal stimulus of high intensity be capable of "burning out" a tinnitus-producing spectral region? (The idea also occurred to Jones and Knudsen [1928].) It would be difficult indeed to enumerate the implicit assumptions underlying such a suggested treatment for tinnitus, and the likelihood is that, once enumerated, the prospects for success would remain sufficiently dubious that one would be reluctant to attempt the procedure even on well-informed volunteers. Fortunately, this difficult ethical predicament need not be faced by anyone, for a bold experimenter has already attempted the procedure on his own tinnitus.

I. M. Young had a unilateral, high-frequency tinnitus that was highly annoying; over a period of time, he administered various high-intensity sounds to himself and then carefully documented the aftereffects of these exposures (Young and Lowry, 1981, 1982). The upshot is that, even after heroic attempts, he has not been successful in eliminating his tinnitus, although he has produced some interesting effects that are difficult to explain. For example, following one series of monaural exposures, the previously monaural tinnitus became permanently binaural. Then, following a monaural exposure of 121 dB SPL at 500 Hz for 21 minutes, the tinnitus frequency was greatly reduced in both ears and it remained so long after hearing had returned to normal. Over the course of weeks, it gradually returned to its original frequency, but with different recovery rates in the two ears.

From this one well-studied subject, it appears that tinnitus will not succumb to high-intensity sounds, but logically some form(s) of tinnitus of different etiology might be reduced by this procedure. Since exposure to intense sounds is known to produce both temporary and permanant tinnitus (as discussed in "Possible Experimental Models of Tinnitus" in Chapter 2), future investigators are encouraged to proceed cautiously even if using themselves as subjects.

Drugs and Tinnitus

Some drugs can cause or exacerbate tinnitus—and thus their termination can alleviate tinnitus. Other drugs can reduce or eliminate tinnitus—and thus are administered to alleviate tinnitus. Since termination and administration of drugs can both be considered forms of treatment, we discuss the two types of drugs here, in successive sections.

Drugs Causing Tinnitus

It is not uncommon for patients with various maladies to report tinnitus for the first time soon after beginning a new drug regimen. For any particular drug, the fraction of patients reporting new or enhanced tinnitus is typically small, but this should not lead us to disbelieve the reality of the reports. Given the large number of ways and locations in which tinnitus can arise, and given the large individual differences in reaction to drugs at the same dosage level, great heterogeneity of response ought to be the rule. Some of the drugs noted for their ability to induce or enhance tinnitus have been noted by Brown et al. (1981) and Goodey (1981). On the other hand, lists such as these and the one prepared by Drucker (1979) must necessarily exaggerate the number of drugs that actually produce tinnitus in a significant fraction of their users. It must be remembered that tinnitus is a very common experience and thus might begin "spontaneously" while someone is taking a drug that should not itself be held responsible for the tinnitus. Considerable work will have to be done before we truly know which drugs possess the ability to induce tinnitus in normal and abnormal auditory systems.


Aspirin, of course, is the most commonly used drug known for its effects on hearing and tinnitus. After just 48 hours on a dosage of about 4.8 g/day, there is 10–15 dB of hearing loss (McCabe and Dey, 1965), and this can grow with continued use to as much as 40–50 dB (Myers and Bernstein, 1965). Typically, the hearing loss is essentially flat across frequency, but in some reports the high frequencies are more affected than the low (McCabe and Dey, 1965; McFadden and Plattsmier, 1982c). Upon termination of the drug, the hearing loss invariably recovers within 24–72 hours, depending upon the serum salicylate level achieved (Myers and Bernstein, 1965; McFadden and Plattsmier, 1982c).

At present there is contradictory evidence as to the site of action of salicylate in the peripheral auditory system. Gold and Wilpizeski (1966) observed diminutions in the whole-nerve response with weak acoustic stimuli but not with intense stimuli (unfortunately, they did not simultaneously record the cochlear microphonic). Silverstein et al. (1967) reported a strong depression in both the cochlear microphonic (CM) and the whole-nerve action potential (N1) following intraperitoneal injections of sodium salicylate. A possible explanation of these two reports is that the salicylate affects the hair cells and that this effect is then reflected in the response of the eight-nerve fibers. However, Wilpizeski and Tanaka (1967) and Mitchell et al. (1973) found no change in CM response, only a diminution in N1 following salicylate injection. Knowing whether or not the CM response is affected by salicylates is basic to insights about site(s) and mode(s) of action of this drug. Resolution of this discrepancy and integration of the outcome with the established effects of salicylates on the stria vascularis (see Woodford et al., 1978, for a review) would be welcome and valuable.

Tinnitus is commonly reported as a concomitant of heavy aspirin usage. Indeed, its presence is routinely used as a gauge of whether the salicylate level in the blood has reached the value believed to be necessary for effective anti-inflammatory action. Thus, dosage is often increased until tinnitus is produced and then decreased until the tinnitus just disappears. (Mongan et al. [1973] warn that this procedure is very unreliable as a predictor of serum concentrations of salicylate, particularly in patients with preexisting, high-frequency hearing loss; many of these people never experience tinnitus, even at very high serum concentration levels.)

Aspirin-induced tinnitus is typically characterized as tonal and of high frequency, but no systematic investigations of its quality, pitch, and loudness as a function of blood salicylate level have been done. McCabe and Dey (1965) apparently did some informal pitch matching, which indicated tinnitus frequencies from 7 to 9 kHz. It has been reported that subjectively the tinnitus onset appears to precede the hearing-loss onset (Mongan et al., 1973), and there is an implication that it also recedes more rapidly than the hearing loss upon termination of the drug (in accord with the personal experience of the author, but see CIBA Foundation, 1981:130–131), but these timing relations are not well documented. Future experiments aimed at this timing issue should incorporate measures of relative hearing sensitivity at frequencies above the normal audiometric range.

No mention could be found in the literature of two features of aspirin-induced tinnitus that the author has personally observed. Attempts to match the pitch of an aspirin-induced tinnitus were frustrated by residual inhibition; a brief, relatively weak matching tone presented to either ear had the ability to eliminate the tinnitus for periods of 30–60 seconds as soon as its frequency approximated that of the tinnitus. Second, clenching of the jaw muscles would produce transient increases in the magnitude of the tinnitus (also see CIBA Foundation, 1981:202; Shulman, 1981a:184). Both of these effects were observed in three listeners taking approximately 5 g of aspirin per day. Unfortunately, the two effects do little to reduce the number of possible explanations of the origin of the aspirin-induced tinnitus. An intriguing but unlikely possibility is that this tinnitus might have as its basis a spontaneous otoacoustic emission (OAE; see "The Objective/Subjective Issue" in Chapter 2).

Great individual differences in serum salicylate levels and hearing loss have been reported for the same weight-corrected dose levels (Myers and Bernstein, 1965; Mongan et al., 1973). Of both theoretical and practical importance would be information on the relationship between susceptibility to aspirin-induced hearing loss and tinnitus and susceptibility to temporary and permanent threshold shift. If eye color and melanin content (see "Lidocaine" in this chapter) are highly predictive of both, an important insight will be gained.

Finally, in regard to temporary threshold shift, Hamernik and Henderson (1976) have reported negative results from animal (chinchilla) experiments aimed at detecting potentiation of noise exposure by administration of sodium salicylate, but McCabe and Dey (1965) reported greater losses in absolute sensitivity following noise exposure when their (human) listeners were on a therapeutic dose of aspirin than when not (in accord with McFadden and Plattsmier, 1982c). This issue should be resolved, given the great quantities of aspirin consumed daily by workers in noisy environments.


Quinine and other antimalarial drugs (e.g., quinidine, chloroquine, and hydroxychloroquine) have long been known for their ability to produce temporary hearing loss and tinnitus (e.g., Brummett, 1980; Brown et al., 1981). With the decline of malaria in the United States, however, these drugs have become infrequent sources of tinnitus in this country, and as a consequence little has been written about this form of tinnitus. Quinine-induced hearing loss and tinnitus are apparently of high frequency, and the tinnitus onset reportedly precedes the hearing loss. While the hearing loss is reported to be reversible in most cases, permanent loss has occurred—apparently following large doses or protracted administration. There appears to be a general belief that the mechanism underlying the hearing loss and the tinnitus is vasoconstriction, although the evidence is meager. Some reports indicate that in sensitive individuals the quinine contained in a single gin-and-tonic highball can be adequate to produce tinnitus.

The occurrence of tinnitus following administration of quinine sulfate is mentioned by Markham et al. (1967) and Segal et al. (1974), and the danger to hearing of chloroquine phosphate is highlighted by the report of Dwivedi and Mehra (1978). Chloroquine was shown to have an affinity for melanin by Dencker et al. (1973), which makes that compound, at least, different from salicylate.


Fowler (1942) asserted that smoking is a common cause of tinnitus and that at least a month's cessation is necessary to eliminate it as a causative factor. Whether or not Fowler was correct in this belief has yet to be satisfactorily established. Tyler (CIBA Foundation, 1981:235) indicates that new information on tobacco is forthcoming.


This agent is frequently mentioned for its ability to produce or exacerbate tinnitus (e.g., CIBA Foundation, 1981:235, 263), but no systematic studies of it were found.


Alcohol has the curious characteristic of being cited as both a cause and a treatment for tinnitus (see ''Alcohol" in this chapter). Unfortunately, anecdotes are the primary source of this information at this time (e.g., CIBA Foundation, 1981:201,268).


Tinnitus is sometimes mentioned as a concomitant to cocaine use, and its vasoconstrictive actions make this claim believable. However, no information was found on the dose levels needed, the time course of onset and decline, etc.


It has been asserted that marijuana can markedly increase a preexisting tinnitus (CIBA Foundation, 1981:168), but no quantitative information appears to exist.

Oral Contraceptives

Some oral contraceptives can produce a hearing loss and an associated tinnitus (Brown et al., 1981). The effects are thought to be due to vascular changes. Detailed information was not found.

Heavy Metals

Tinnitus is a common side-effect of heavy-metal treatment for cancer (Merrin, 1978; Merrin et al., 1978; Von Hoff et al., 1979). For cis -platinum at least, the evidence indicates that the symptom is reversed upon withdrawal of the drug.

Drug Therapy for Tinnitus

A wide variety of drugs have been investigated as possible treatments for the symptom of tinnitus. The basis for interest in a particular drug or class of drugs is frequently not explicitly stated by the investigator, which suggests that it came to attention simply because patients receiving the drug for some other medical problem reported their tinnitus to be alleviated. Given the general level of ignorance about the sites of origin of and the mechanisms underlying tinnitus, this empirical origin of ideas about tinnitus-alleviating drugs should not surprise us. The fact is that greater success has been achieved with drugs discovered in this way than with drugs studied because of some theoretical view about the origins of tinnitus.

The list of drugs investigated at one time or another as a possible therapy against tinnitus is long (see Goodey, 1981), and no attempt is made here to be exhaustive. There are some general points about experimental design that the reader ought to keep in mind while reading the following synopses of the studies done on drug treatment of tinnitus.


Because tinnitus is not a single entity produced by a single or even a small number of causes, one should not expect any one drug to be universally effective against tinnitus. Further, the actual proportion of the sample of patients reporting relief in a particular study will depend heavily upon the makeup of that sample. If, for some reason, an investigator ended up with a sample heavily dominated by patients whose tinnitus arose from the same problem—(say) in the cochlea itself—and if the drug being studied proved to be extraordinarily effective against that problem, that investigator might have a very high success rate, while another investigator doing the same manipulations, but on a sample less heavily dominated by patients with that cochlear anomaly, might have a much lower success rate. Even in those experiments in which the audiological diagnoses on the patients are reported, one cannot be confident about homogeneity in the site(s) of origin of the tinnitus.


The severity of the tinnitus symptom varies greatly across the patients used in different studies and sometimes across those in the same study. In some studies, the patients used are referred to the investigator simply because they have tinnitus, of whatever magnitude; in other studies, only severely afflicted patients are used as subjects. Comparison of the success rates across studies as different as these is problematic enough, let alone trying to generalize to the population of tinnitus sufferers at large.


Many of the drug studies of tinnitus do not include the controls usually regarded as necessary for unambiguous interpretation of an outcome. For example, few studies utilize single-or double-blind procedures, control and experimental groups, or cross-over designs.


Because of 2 and 3 above—the great heterogeneity in the makeup of the patients in different drug experiments and the absence of typical control procedures—it is difficult to evaluate the contribution of the so-called Hawthorne effect (Roethlisberger and Dickson, 1939) to the reported success rates. When some of the patients have been through months or years of frustration trying to get relief from their tinnitus, it may be that anyone's doing anything in an attempt to help them might be met with reports of success.


It is very rare in drug studies of tinnitus for an investigator to manipulate the dose level or the duration of the administration, either within or across subjects. Thus, the reader is always uncertain whether some of the patients in the "no relief" category might have moved to the "partial relief" category and/or some of the "partial relief'' to the "complete relief" category had the dosage or the period of administration been changed.


Drug studies of tinnitus often do not include monitoring of, or reporting of, drug serum levels. This is an understandable omission in many cases, since the taking and analysis of numerous blood samples raises complications for both the patient and the experimenter. Without such information, however, it is impossible to know whether the difference between success and failure of treatment in individual patients might be due to differential compliance with the dose schedule and/or differential absorption of the drug, for whatever reasons. It is possible that some of the across-subject variability in effectiveness of some drugs may be traced to different "effective doses" reaching the relevant site, and while drug serum level is an imperfect control for this problem, it does constitute a necessary first step. A procedure that seems sensible is one in which fixed or weight-corrected doses are given to sufferers from a wide variety of tinnitus types in the early experiments on a particular drug, and then—given that there is some evidence of effectiveness of the drug—the serum levels of a subset of the therapeutic successes and failures might be checked in a follow-up study.


A member of the vitamin B complex, niacin has three common forms—niacinamide, nicotinic acid, and nicotinamide. The amino acid tryptophan can also be converted into niacin by the body. Niacin is a peripheral vasodilator and thus has been used in the treatment of peripheral vascular disorders and migraine headaches. It also has a long history in the treatment of some forms of Meniere's Disease. Atkinson (1944a) found that about 50 percent of his (selected) patients reported relatively long-term relief from their tinnitus while under chronic maintenance dosages of nicotinic acid—injections initially, then oral administration. Flottorp and Wille (1955) used balancing and masking procedures to estimate the magnitude of tinnitus before and during administration of daily nicotinic acid (again, injection initially, oral eventually). As with the Atkinson study, most of the Meniere's patients reported an improvement in their tinnitus. In addition, nearly all of the patients in a group having normal or near-normal hearing evidenced lower matching intensities as well as subjective improvement. Subsequent experience with niacin treatment has not been as positive, so while it is still occasionally used, it is not a routine component in the treatment of Meniere's Disease or of tinnitus in general.

Vitamin A

Graham (1965:Table 4) provides a summary of the studies concerned with vitamin A therapy. The early reports of both reductions in tinnitus and improvements in hearing sensitivity following massive intramuscular injection of vitamin A were not confirmed in later studies, and the issue seems to have been dropped. It is possible that people with dietary deficiencies of vitamin A do, in fact, realize both improvement in their hearing and reduction of tinnitus but that people with more normal diets cannot profit from increased intake of vitamin A. (A similar conclusion was reached by Ward and Glorig [1960] in a study of vitamin A and temporary threshold shift.) Because it is possible to overdose on vitamin A, it is not advisable for tinnitus sufferers to experiment with heavy doses of this agent without medical supervision.


(also known as lignocaine; Xylocaine is a registered trademark) Several recent reports indicate a high rate of improvement in tinnitus following administration of lidocaine. This amide is commonly used as a local anesthetic in surgery of the middle ear and upper respiratory tract, but it is also a potent short-term anticonvulsant that has vasodilation as one of its effects. During routine systemic administration of this drug for controlling and isolating the origin of central pain, it was noticed that preexisting tinnitus was alleviated or eliminated. (Interestingly, a single administration has sometimes alleviated the tinnitus for up to 72 hours.) Subsequent research has substantiated its short-term effectiveness against long-standing, intractable tinnitus. Lidocaine itself is not suitable for use in a tinnitus management regimen because it must be administered intravenously, and it has a short half-life in the body. However, its use as a research tool has stimulated much interest in drug treatment for tinnitus, and, as a consequence, more suitable compounds are being developed and tested (e.g., see tocainide and carbamazepine below).

Gejrot (1963) may have been the first to note that lidocaine can affect tinnitus. He injected the drug during episodes of Meniere's Disease and found, among other things, that the tinnitus disappeared for about 20 minutes in all 11 patients described.

Melding et al. (1978) studied 78 patients with "incurable and intolerable" tinnitus. Prior to the single administration of lidocaine, each patient was given a standard audiological examination. Following administration (I.V. injection of 1–2 mg/kg over a period of 3–4 minutes), the audiometry was repeated, and the patients were asked to indicate any improvement in tinnitus by placing a mark along a 100-mm line to indicate what fraction of the original tinnitus remained. (Exactly when following the injection this judgment was made is not specified, but presumably it was within a few minutes of injection.) Approximately 35 percent of these severely affected patients reported total abolition of the tinnitus, and an additional 28 percent reported the tinnitus to be 0.3 or less of its initial magnitude; about 26 percent reported no relief from their tinnitus. This has to be viewed as a great success, given the ordinary refractoriness of severe tinnitus. The authors report that the relief "lasted between 10 minutes and three days," but without specification of (1) how the data for the longer intervals were collected or (2) how the return of tinnitus progressed. These points are important, given the high likelihood that different patients were using different criteria for their judgments.

Some of the patients reporting relief from tinnitus claimed that it was accompanied by an improvement in hearing. There is as yet no evidence to support this impression; the postinjection pure-tone audiometry did differ in some of these subjects, but the differences appear to be within the typical test-retest range of variation. Nevertheless, claims of improved hearing are reasonably common in drug studies on tinnitus (see the discussion of Melding and Goodey [1979]; and Shea and Harell [1978] in the next section). It may be that these patients' subjective impressions of improved hearing stem solely from the elimination of the ordinarily present and distracting tinnitus, but it is worth investigating whether or not some aspect of auditory function other than pure-tone sensitivity—usable dynamic range, for example—is altered by lidocaine.

When Melding et al. (1978) partitioned their 78 patients into standard diagnostic categories, they discovered that those with sensorineural problems had the greatest chance of benefitting from the lidocaine injection. Of the 55 patients with ''pure sensorineural hearing loss," three subgroups—constituting 41 patients—stand out, as shown in the following breakdown of patients reporting either a total abolition of tinnitus (excellent relief) or a reduction in it to 0.3 or less of its preinjection magnitude (good relief):

Excellent Relief Good Relief Total
Endolymphatic hydrops4/82/875%
Acoustic trauma/noiseinduced hearing loss9/164/1681%
Presbycusic hearing loss11/174/1788%

These samples are small and, as indicated above, highly selected for severity of tinnitus, but the suggestion appears strong that short-term relief is possible for the majority of people with sensorineural-based tinnitus. The mechanism(s) mediating this improvement remain unknown, but a valuable tool seems to have been added to the scientists' chests. (For additional lidocaine results, see the following section on carbamazepine.)

An exception exists to the general criticism about the lack of appropriate controls in drug studies of tinnitus; Martin and Colman (1980) used a double-blind procedure in which each subject served as his or her own control (some-times known as a cross-over design). The subjects were patients referred by otologists because of their tinnitus, but apparently they were not as highly selected for severity of tinnitus as in other studies. They were given standard physical and audiometric exams, they matched their tinnitus in frequency and intensity, and their tinnitus was rated (by the experimenters) on a three-point scale of severity. At random, the subjects were given an intravenous dosage of lidocaine (1.5 mg/kg of body weight of a 2 percent solution) or of an equal volume of normal saline; then they again matched their tinnitus, had it rated, and indicated how much the tinnitus had been relieved. After a time interval corresponding to "more than double the physiological half-life of intravenous lidocaine" (presumably more than 200 minutes total), the second substance was injected and identical follow-up measures taken.

In the lidocaine condition, there were both objective and subjective improvements in tinnitus. Objectively, the lidocaine injection produced reductions in both the frequency and the intensity of the sounds chosen as matches to the tinnitus. The decreases in matched frequency appear to be real, but small; no explanation of these changes is provided or comes to mind. Following lidocaine, about 31 percent of the subjects matched their tinnitus to intensities smaller than those before injection by 7.5 dB or more (versus 6 percent following the saline injection); an additional 62 percent showed little or no change following lidocaine (versus 69 percent following saline). This is one of the few studies we know reporting an objective change in tinnitus following treatment (compare Donaldson [1978], discussed in "Sodium Amylobarbitone" in this chapter); the outcome is important and deserves further documentation, including study of its time course.

The subjective results in the Martin and Colman study are even more impressive than the objective results. One would expect that the reductions in the matched intensity of the tinnitus would produce concomitant reductions in the perceptibility or the annoyance of the tinnitus, and, in fact, 72 percent of the subjects judged their relief from tinnitus to be 50 percent or greater following the lidocaine injection (versus 9 percent following saline). This percentage is similar to the 61 percent reporting good or excellent relief in the Melding et al. (1978) study. The percentage reporting total abolition of tinnitus following lidocaine was 35 percent in the Melding et al. study and 44 percent in the Martin and Colman study (recall that the former study used only subjects afflicted with severe tinnitus). Unfortunately, Martin and Colman provide no information on the time course of this temporary relief; such information could prove valuable.

Thus, Martin and Colman's use of a saline placebo seems to strengthen the case of I.V. lidocaine as an effective agent against tinnitus, and it is to be hoped that similar research designs will be adopted by other experimenters working in this area.

Emmett and Shea (1980) briefly reported the results of a study that points to a curious possibility. Intravenous lidocaine was administered to 592 patients having "extremely bothersome tinnitus." One minute after injection, the patients were asked to indicate the magnitude of tinnitus relief in percentage units. Unlike other studies, the data are reported separately for the unilateral and the bilateral tinnitus sufferers. It is interesting that the bilaterally afflicted were much more likely to report relief than were the unilaterally afflicted. About 73 percent (139/190) of those with bilateral tinnitus reported relief of 50 percent or greater from the lidocaine injection, while only about 49 percent (197/402) of those with unilateral tinnitus reported relief in that range. At the other extreme, no relief was reported by only about 6 percent of those with bilateral tinnitus and about 27 percent of those with unilateral tinnitus. This possible difference in lidocaine effectiveness against unilateral and bilateral tinnitus has both theoretical and practical importance, and thus it is necessary to verify it using tinnitus matching procedures.

The mechanism(s) underlying the effectiveness of lidocaine on tinnitus are being studied. Using peripheral nerve preparations, Condouris (1976) has shown that the primary effect of local anesthetics such as lidocaine is to prolong a neuron's refractory period. This in turn produces a reduced ability to transmit high rates of firing. The intuitive view of how local anesthetics act to reduce pain is thus confirmed; firing rate declines, and this is accompanied by a diminution in perceived pain. By implication, then, what lidocaine may accomplish in regard to tinnitus is a reduction in the rate of activation of some (primary?) auditory fiber(s) sending a misleading signal.

Supplementary information on mechanisms comes from other research. Comeau et al. (1973) showed that iontophoresis of a lidocaine solution from the middle ear cavity into the cochlea did not alter cochlear microphonic potentials recorded for frequencies from 100 Hz to 20,000 Hz. However, Hughes and Yagi (1975) reported that lidocaine injected into the middle ear cavity produced a "prompt and prolonged" depression in the cochlear microphonic, and Yagi et al. (1978) noted pronounced reduction in resting rate and in responsivity of primary vestibular neurons of cats. Gerke et al. (1977) have provided evidence that lidocaine blocks the uptake of adrenalin (but not noradrenalin) by sympathetic nerves surrounding the media of certain arteries. Rahm et al. (1962) warn that lidocaine hydrochloride alone can have highly unpredictable cochlear effects across individual animals, but that when mixed with a vasoconstrictor such as epinephrine, the effects are more homogeneous. (The commercial form of lidocaine—Xylocaine—can be obtained with or without epinephrine added.)

Perhaps the most interesting insights about the mechanisms underlying lidocaine action come from Lyttkens and his colleagues (1979a,b). They have reported that lidocaine (as well as bupivacaine and chlorpromazine) bind to melanin in the inner ear. Lyttkens et al. believe that the melanin is involved in energy transformation—mechanical to electrical—and that in some way lidocaine may act as an aid to this process in a defective ear, thereby reducing the tinnitus. Whatever the final explanation, the findings of Lyttkens et al. raise several interesting points regarding drug treatment of tinnitus.

This relationship between lidocaine and melanin is intriguing, for it is now well established that melanin content is strongly related to a person's susceptibility to both temporary (Hood et al., 1976) and permanent (Karsai et al., 1972) hearing loss due to noise exposure—the more heavily pigmented a person is, the greater is the resistance to hearing loss. One interpretation of this fact is that the melanin in the inner ear is in some way contributing to the ear's line of defense against damage from overstimulation. The demonstration by Lyttkens et al. that an agent apparently quite effective against some forms of tinnitus binds to the melanin in the inner ear raises the two questions of whether tinnitus incidence is also correlated with melanin content, and whether dose levels of lidocaine (and some other tinnitus-ameliorating agents) might better be tied to skin or eye color than to body weight. It would be helpful to know whether a particular skin or eye color is disproportionately represented in the groups showing weak or no responses to lidocaine or carbamazepine (see following section). In a discussion at a recent conference (CIBA Foundation, 1981:27–28), several clinicians agreed that dark-skinned people were greatly underrepresented in their practices, and Coles indicated that the recent British survey of tinnitus did include eye color and race data that are now being analyzed. (W. F. House [1975] believes that Meniere's Disease is less prevalent in blacks than in whites, and, if this proves to be true, it may also be related to inner-ear melanin.) Melanin content appears to be a variable highly worthy of attention in future surveys and experimental studies of tinnitus.


(Tegretol is a registered trademark) This drug is an oral anticonvulsant and mild antidepressant that is best known in this country for its effective use on trigeminal and glossopharyngeal neuralgia; in Europe it has been used extensively against epilepsy.

Based on their view of the origin(s) of tinnitus—"an abnormal hyperactivity"—Melding and Goodey (1979) postulated that carbamazepine might be effective against tinnitus. As in the Melding et al. (1978) study discussed in the section on lidocaine, only patients with "incurable and intolerable" tinnitus were included in this study. As in the earlier work, a full audiological exam was performed prior to any drug treatment. Then lidocaine was administered and its effectiveness measured as described above. The intent of this phase of the procedure was to obtain the data necessary for later correlation of the effectiveness of lidocaine and carbamazepine. (It is not clear from the published report whether the outcome of this phase was known to the experimenters when conducting subsequent interviews and measurements and/or whether the patients were asked to, or dissuaded from, comparing the effectiveness of lidocaine with that of the carbamazepine they later received.) As an aside, it should be noted that the effectiveness of lidocaine was about What it was in the previous experiment—about 77 percent of the patients reported total abolition of their tinnitus or a reduction in its magnitude to 0.3 or less of its pre-injection value (although it is not clear whether this group of 125 patients included some or all of the 78 patients reported on in the previous experiment).

Regardless of how they fared on the lidocaine injection, all the patients studied by Melding and Goodey were offered treatment on carbamazepine. Dosage began at 100 mg three times daily and was gradually increased until therapeutic levels were reached (typically 600–1000 mg/day); dosage was then stabilized, and treatment continued for 2–3 months. Once dosage was stabilized, the patients scaled any improvements in tinnitus by making a mark along a 100-mm line. Of the patients reporting total abolition of tinnitus following a single I.V. injection of lidocaine, about 60 percent reported either total abolition of tinnitus or reduction in its magnitude to 0.3 of its original magnitude or less by the carbamazepine regimen. Of those reporting major reduction but not abolition of tinnitus from lidocaine, about 53 percent reported good-to-excellent relief from carbamazepine. In general, no patient for whom lidocaine failed to produce relief received relief from the carbamazepine regimen (but see below). As before (Melding et al., 1978), no significant changes in audibility could be measured even in those patients reporting an apparent improvement in hearing (see comment under "Lidocaine" in this chapter). Once the carbamazepine was withdrawn, the tinnitus returned to its original magnitude within 2–3 weeks.

Shea and Harell (1978) followed up on the Melding and Goodey demonstration with lidocaine and carbamazepine. Their patients were also highly selected for severe and intractable tinnitus and, again, about 80 percent of these reported significant improvement in their tinnitus following a single injection of lidocaine (the procedure for measuring tinnitus was not specified). Unlike the Melding and Goodey procedure, however, in this study only those patients showing relief from the lidocaine injection were allowed access to the carbamazepine regimen (27 patients); this screening was later shown to be of questionable value. Shea and Harell also used lower daily dosages of carbamazepine than did Melding and Goodey (100–400 mg/day in contrast with 600–1000 mg/day), with little apparent effect on the success rate; approximately 80 percent of the lidocaine-screened sample reported partial or complete tinnitus relief from carbamazepine (the same data are reported by Emmett and Shea [1980]). This is an encouraging finding, since the incidence of objectionable side-effects from carbamazepine is dose-related. Again, some patients reported a subjective improvement in hearing that could not be confirmed with the standard audiometric tests. Finally, Shea and Harell report one case of particularly severe palatal myoclonus (rhythmic, rapid contractions of the soft palate) that succumbed to a daily regimen of a low dosage of carbamazepine.

An important outcome of the Shea and Harell (1978) experiment was that several patients who barely qualified for the carbamazepine phase, due to minimal response to the lidocaine injection, did profit from carbamazepine. This—plus the fact that some patients receiving excellent relief from lidocaine received only minimal relief from carbamazepine—seems to make screening on the basis of the lidocaine injection a questionable procedure. Given the current level of understanding about the actions of these drugs, future investigators may wish to consider eliminating such screening (but note that Goodey [1981] disagrees).

It must be noted that carbamazepine can produce a wide variety of side effects (Goodey, 1981) and carries a significant risk of hepatotoxicity; thus, careful monitoring is required of patients taking it. This obviously limits the usefulness of carbamazepine as a general treatment for severe tinnitus, but the experience of Shea and Harell shows that the problems may not be insurmountable.

Tocainide Hydrochloride

(Tocard is a registered trademark) Unlike lidocaine itself, which must be administered intravenously, this analog of lidocaine can be taken orally because of a difference in its metabolism by the liver. Additional attractions are that it has a physiological half-life of about 11 hours as compared to about 1.5 hours for lidocaine, and it has fewer side effects. Tocainide is currently being used experimentally as an oral anti-arrhythmic agent in cardiac patients.

Two reports exist of attempts to treat tinnitus with tocainide (Emmett and Shea, 1980; no date); since the former consists primarily of preliminary results of dose level, we shall discuss only the latter (also reviewed by Emmett, 1981). Subjects not showing at least 75 percent relief of tinnitus following I.V. administration of lidocaine were screened out; the remainder were assigned to one of three groups: 600 mg of tocainide at 6-hour intervals four times a day; that same dose or a placebo every other 6-hour interval (i.e., tocainide twice a day), or a placebo four times a day. The duration of the study was 2.5 days (10 dose intervals); it was double blind.

Of the 56 patients beginning the study, 11 (20 percent) dropped out because of nausea (two of these were receiving a placebo only). Ninety minutes following the last dose, the patients rated their relief from tinnitus. Relief was judged to be between 40 percent and 100 percent in 46 percent (6/13) of the patients receiving four doses/ day, in 41 percent (7/17) of those receiving two doses/ day, and in 7 percent (1/15) of those receiving only a placebo.

Emmett and Shea used a self-report procedure in an attempt to determine the duration of relief following termination of the drug. The group that received tocainide four times a day reported a return of the tinnitus to its initial loudness after an average of 59 hours, while the average for the two-a-day group was 20 hours (11 hours for the placebo group). That is, the higher dose level did produce longer-lasting relief even though it did not significantly increase the fraction of patients receiving relief in the first place (46 percent versus 41 percent in the four-a-day and two-a-day groups, respectively). This outcome is intriguing, but requires verification using tinnitus matching procedures. Emmett (1981) noted that some patients who did experience tinnitus relief from lidocaine did not from tocainide, but there is no information about the converse situation, since lidocaine was used to screen the patients for the tocainide trials. The advisability of such screening procedures was discussed under ''Carbamazepine" above.

It appears, then, that tocainide is not as effective in reducing tinnitus as are lidocaine and carbamazopine. On the other hand, it is clear that clinicians and experimentalists do have an additional weapon in their arsenal against tinnitus. Finally, an intriguing possibility was raised by Hazell (CIBA Foundation, 1981:50); he reported that several patients for whom neither tocainide nor masking was effective did benefit from masking when it was combined with tocainide (also see Vernon and Meikle, 1981).

Phenytoin Sodium

(also known as diphenylhydantoin; Dilantin is a registered trademark) This drug is an oral anticonvulsant similar in action to carbamazepine. In the Melding and Goodey (1979) paper discussed above, brief mention is also made of Dilantin. It was apparently used on four patients who developed an allergic reaction to carbamazepine, but the only indication as to how these patients fared is the comment that Dilantin is always less effective than carbamazepine in treating tinnitus (Goodey, 1981). Shea and Harell (1978) reported that none of 15 patients treated with this drug had any relief from tinnitus, even though all had some relief from the lidocaine injection.


(Mysoline is a registered trademark) This anticonvulsant has been tried against trigeminal neuralgia with mixed results. Emmett and Shea (1980) briefly mention a study of its effectiveness against tinnitus. Patients initially received 250 mg twice daily; the dosage was increased monthly in increments of 250 mg/day up to a maximum of 2 g/day or until the tinnitus was relieved. Details are few, but apparently 27 percent (11/41) of the patients reported 80–100 percent relief, and an additional 59 percent (24/41) reported 20–80 percent relief from their tinnitus. Exactly when in the regimen these judgments were made is not revealed, yet this appears to be an important issue given the high incidence of side effects reported.

Sodium Fluoridi

This compound has come to be recognized for its ability to reverse the process of demineralization in the cochlear capsule that leads to the condition of otospongiosis. Since this condition is often accompanied by a sensorineural-type hearing loss and by tinnitus and vertigo, it is of theoretical as well as practical interest here that these latter symptoms are often diminished or abolished as the otospongiosis is reversed (Shambaugh, 1977). The explanation offered is that the otospongiotic focus gives off cytotoxic enzymes that then enter the perilymph, causing a deterioration of cochlear elements critical to the normal transduction process , and, as a side effect, producing a concomitant tinnitus. The administration of sodium fluoride cannot reverse the hearing loss (although according to Shambaugh it does arrest it), but it can eliminate the tinnitus and vertigo. The process of reversal of bone damage often takes several years of daily doses of 40 mg of sodium fluoride. No information about the time course of reversal of the tinnitus is given, nor apparently were any psychophysical measures of tinnitus taken.

Sodium Valproate

(Depakene, Epilim, and Ergenyl are registered trademarks) Goodey (1981) briefly mentions this drug, indicating that it reduces tinnitus in about the same proportion of subjects as does carbamazopine, but that, at the doses used to date, the amount of relief experienced is less than with carbamazepine. Apparently side effects are minor.

Sodium Amylobarbitone

Noting a relationship between drugs that are effective on trigeminal neuralgia and on tinnitus (e.g., carbamazepine), Donaldson (1978) selected this fast-acting barbiturate, known for its effectiveness on trigeminal neuralgia, for a study on tinnitus. Forty patients with tinnitus of varying severity were randomly assigned to the experimental or control groups. Prior to treatment, all patients were assessed audiometrically, were asked to match the pitch and loudness of their tinnitus, and were asked to rate their tinnitus on a four-point scale (from "only noticeable in quiet environments" to "interferes with sleep, and patient engages in some activity to distract attention from it"). The experimental group was then put on a regimen of 50 mg in the morning, 50 mg in the early afternoon, and 80 mg at night; tinnitus was reassessed after 6 and 12 weeks. The drug was withdrawn after 12 weeks, and a final assessment was made at 18 weeks. Apparently the control group received absolutely nothing, but these subjects were reassessed at 6 and 12 (but not 18) weeks. Thus, the study was not "blind,'' and the control group was primarily a control for the passage of time (and for the spontaneous remissions that might occur in that time). Prior to treatment, the majority (70 percent) of both the experimental and control groups matched their tinnitus to frequencies of 400–10,000 Hz, and 78 percent matched to intensities of 20 dB SL or less (in good accord with Reed, 1960). After 12 weeks, the control group changed very little on these measures, but marked changes occurred in the experimental group. Thirteen (of 20) in the experimental group (compared with 2 in the control group) matched to lower intensities—this included 4 subjects for whom the tinnitus was abolished—and many now matched to a lower frequency, a fact that stands unexplained (but was also obtained by Martin and Colman, 1980). Perhaps most important is the fact that, of the 12 experimental subjects who, prior to treatment, rated their tinnitus as being in one of the two most severe categories, only 1 persisted in this rating at the 12-week reassessment. That is, 11 of the 12 most severe cases (92 percent) benefitted from the amylobarbitone regimen, and, for 2 of these, the tinnitus was abolished.

Thus, Donaldson's study provides evidence that sodium amylobarbitone may be quite effective in diminishing the disturbance caused by tinnitus. Its usefulness as a treatment for episodic or chronic tinnitus is somewhat limited by its potential for damaging the liver, but in the (apparently effective) dosage used by Donaldson, no such problems arose. One curious aspect of Donaldson's report is that at the 18-week reassessment—6 weeks after drug usage had ceased—there were "no significant alteration(s)" in the states of the experimental subjects, including no return of tinnitus in the four subjects who had reported it abolished. This implies a reversal of the conditions initially responsible for the tinnitus—an unlikely event—which makes this an outcome requiring verification through further study.


Excess alcohol consumption is frequently mentioned in anecdotes as a cause of tinnitus, but curiously, some tinnitus sufferers apparently find that small amounts of alcohol assist in tinnitus control (Goodey, 1981). Of theoretical and practical interest is the nature of the assistance. If the primary contribution of the alcohol is to general relaxation and improved coping ability, that is far different from an actual physiological effect on tinnitus magnitude. Psychophysical measurements are in order. However, the fact that large numbers of tinnitus sufferers have not independently discovered palliative effects in alcohol consumption implies that psychophysical measurements will fail to establish effects on the tinnitus itself.

Miscellaneous Drugs

Here we list drugs that have come to our attention, but about which little is known.

Hazell and Jackson (CIBA Foundation, 1981:278) mention informal studies using naftidrofuryl, and Jackson (CIBA Foundation, 1981:277) mentions lorcainide and bupivacaine.

Arlidin, a vasodilator, and chlortrimeton, an antihistamine, have been used singly and in combination with some success against some forms of tinnitus (Shulman, 1981b).

Goodey (1981) indicates that the tricyclics, especially nortriptyline, may be worth investigating as tinnitus-reducing agents.

Diazepam (Valium) is surely one of the most widely prescibed drugs for tinnitus and its psychological concomitants, but Goodey (1981) says that there is no evidence that it has any value in this regard, and it can make depressed patients worse.

McCormick and Thomas (1980) utilized a double-blind cross-over design in a study of mexilitine—a pharmacological relative of lidocaine that has an oral form. The patients' numerical estimates of the severity of their tinnitus were unaffected both by this drug and by the placebo.

It has been asserted that barbiturates do not cause tinnitus (CIBA Foundation, 1981:169; see "Sodium Amylobarbitone" in this chapter), and it has been suggested that the aminoglycosides may produce only a temporary tinnitus (CIBA Foundation, 1981:34).

Heparin is reported to have produced temporary relief from tinnitus in a number of heart patients (CIBA Foundation, 1981:170).

Trowbridge (1949) injected a 5 percent solution of ethylmorphine hydrochloride—an analgesic and vasodilator—directly through the tympanic membranes of patients suffering from tinnitus that he judged to be caused by structures of the middle ear (the tympanic plexus). He injected repeatedly at 4-day intervals and claimed reduction or elimination of tinnitus for the majority of patients so treated. He also claimed improved audiometric measures. No recent application of these procedures were found.

Conclusions About Drugs and Tinnitus

It is important to emphasize the particular need for appropriate controls and measurements in studies of drug treatment for tinnitus. The absence of balanced experiments and of blinds and the likely presence of the Hawthorne effect and a "flight to health" are a potentially troublesome mix when the goal is to determine efficacy of a procedure. Nevertheless, recent discoveries about the effectiveness of certain drugs against some forms of tinnitus are encouraging. Unfortunately, lidocaine, carbamazepine, and sodium amylobarbitone are all very potent drugs that do carry possibilities of serious side effects that limit their usefulness; however, there are some indications that relatively low doses may be effective against tinnitus (Shea and Harell, 1978), and related, safer compounds may be identified. We clearly are not yet at the stage of being able to prescribe a safe and effective drug for every tinnitus sufferer, but the past several years have seen the establishment of a firm base of knowledge, and there is every reason to believe that further progress will be made. The effort is definitely worthwhile.

Some intriguing facts coming out of this research deserve verification and further study:


There is an interesting possibility that the mechanism(s) through which lidocaine works against cochlear-originating tinnitus involves melanin, a substance previously linked with susceptibility to noise-induced hearing loss. It is of theoretical interest to know if there is also a greater susceptibility to severe tinnitus among those less well endowed with melanin. Of both theoretical and practical significance are the questions of whether the effectiveness of drug (or other) therapy is related to melanin concentration, and whether susceptibility to the reversible tinnitus induced by drugs and exposure (see "Possible Experimental Models of Tinnitus" in Chapter 2) also varies with melanin concentration.


The drugs found to be effective against severe tinnitus have been reported to have other auditory effects. Many patients report a marked subjective improvement in everyday hearing; to date these impressions have not been confirmed by objective test (Melding et al., 1978; Melding and Goodey, 1979; Shea and Harell, 1978; Emmett and Shea, 1980; CIBA Foundation, 1981:190–191), but it is possible that audiometric measures other than pure-tone sensitivity would reveal improvements. Second, there is some evidence from matching procedures that, following certain drugs, the magnitude of the tinnitus is reduced (Donaldson, 1978; Martin and Colman, 1980), and its frequency is lowered (Donaldson, 1978; Martin and Colman, 1980). The former outcome is in accord with patients' reports of the distress from tinnitus being reduced, but the latter outcome is curious and yet to be explained. Both deserve experimental attention.


Goodey (CIBA Foundation, 1981:288) claims that following long-term drug therapy about one-half the tinnitus sufferers never again experience tinnitus at its previous level. Numerous explanations of this fact spring to mind, but first it is necessary to verify the assertion.


There exist hints that drugs can work in conjunction with masking therapy to produce a greater reduction in tinnitus than either can produce singly (CIBA Foundation, 1981:50).

Allergy and Diet

It is believable that any allergic reaction that directly or indirectly affects the outer-, middle-, or inner-ear systems could also produce a concomitant tinnitus. For example, any food or inhalant allergy that caused a blockage of the Eustachian tube might produce tinnitus as a byproduct. So, in theory, various allergy-induced changes in the inner ear might produce tinnitus, but published evidence of such changes is scanty. Goodey (CIBA Foundation, 1981:234,263) comments on the role diet control plays in tinnitus management and notes that in his experience coffee, tea, tonic water, red wine, grain-based spirits, cheese, and chocolate have been the most common dietary sources of tinnitus. But beyond comments of this sort, nothing is published.

Allergies and their control seem to have been given the greatest attention in regard to Meniere's Disease. Belief persists among some clinicians that allergies, particularly food allergies, are responsible for, or are a significant secondary factor in, some cases of Meniere's Disease, and since this malady is typically accompanied by a low-frequency, roaring tinnitus, this work deserves mention.

There can be little doubt that occasionally a patient will be found who shows marked vestibular or auditory effects as part of a general allergic reaction to certain agents and that control of these agents controls the aftereffects (see Wilson, 1972). The issue is how representative such patients are. Shaver (1975) was only able to relieve the Meniere's symptoms of about one-third of his patients through allergy management, and the success rate of Endicott and Stucker (1977) was apparently even lower. Pulec (1972), Clemis (1975), and Powers (1975) report equally modest success rates. All of these authors comment on the great difficulties and uncertainties associated with isolating the allergy-producing agents in an individual's daily life.

Thus, it appears that allergy management may aid some tinnitus sufferers who also have Meniere's-like symptoms, but the published success rates indicate that unless severe allergies are part of the patient's medical history, allergy management should be a course of last resort or only part of a general plan for managing tinnitus arising from Meniere's Disease (see, for example, Goodhill, 1979:541).


In general, biofeedback procedures are designed to give a subject control of a physiological variable over which he or she previously had no conscious control. Common examples are blood pressure and heart rate. In a typical procedure, the relevant physiological variable is monitored, and changes in the desired direction are indicated (fed back) to the subject via a simple auditory or visual stimulus. The subject's task, of course, is to learn to produce the desired changes through any means available. The goal is to eventually be able to dispense with the monitoring and signaling devices, so that the subject can effect the relevant changes whenever desired or necessary in his or her everyday life.

Biofeedback has been attempted on patients suffering from tinnitus, but in no instance has the prototypical procedure just described been used. That is, the subject has not been trying to learn how to reduce the tinnitus itself; rather, the task has been to learn to "relax"--as defined by electromyographic changes in tension in (for example) the frontalis muscle. There are two reasons for this indirect approach: (1) the obvious one that in most cases the tinnitus has not been detectable by external sensors (human or electronic) and thus could not serve as the basis for a feedback signal, and (2) a line of reasoning that goes: tinnitus magnitude has been reported to co-vary with level of anxiety, anxiety is widely believed to be related to muscular tension, and therefore reduced muscular tension ought to reduce tinnitus (a view that surely oversimplifies the etiology of tinnitus). Apparently, few biofeedback practitioners argue that the tinnitus per se is affected by the biofeedback sessions (P. R. House, 1981, may be an exception), but rather that the (not to be denied) gain in control over general relaxation simply makes the tinnitus more tolerable. This does not invalidate the procedure, of course, but it does put it in a different category from procedures that may, or do, alter the source of the tinnitus. It is possible that a direct biofeedback procedure could be developed for people whose tinnitus has an objective, acoustical basis, so that the patient could try to learn to control tinnitus magnitude in the same way that control of blood pressure and muscle tension are learned. To our knowledge, such a procedure has never been attempted.

Grossan (1976) reports the results of a typical biofeedback experiment. In that experiment, all 11 presbycusic patients "felt that their tinnitus was improved" in the sense that "they could now cope with it"; tinnitus magnitude, however, as measured by a matching procedure, was unchanged following the six 20-minute sessions. Fourteen of 18 patients with acoustic trauma reported some improvement following biofeedback training, but, just as with the presbycusic group, there was no major change in the tinnitus matching levels. Only 2 of 21 patients suffering from cranial or cervical trauma reported some relief from tinnitus, although 13 others did report improvement of neck symptoms.

J. W. House et al. (1979) studied only patients with severe tinnitus who were persistent in their searches for relief but who were "therapeutic failures." After 10–12 one-hour biofeedback sessions spent learning to relax the frontalis muscle and increase peripheral blood flow, about 10 percent of the 41 patients stated that their tinnitus was now absent, and about 80 percent felt that there had been some improvement in their tinnitus. Better sleep and less need for medication (tranquilizers and antidepressants) were also reported. No breakdown of these proportions by type of auditory pathology is provided.

Thus, biofeedback training does appear to offer hope to some tinnitus sufferers, but it is likely that whatever improvement results is due more to the patients' increased ability to cope with the symptom than to its having been reduced. It is possible that direct methods of biofeedback training could be developed in the future and that, with these, true reduction in some forms of tinnitus might be realized. Certainly biofeedback ought to be considered before more psychological treatments, such as formal counseling or psychotherapy, if only because patients are more likely to participate and cooperate.


We know of no large-scale study on the effectiveness of hypnosis on tinnitus. Marlowe (1973) reported on two cases in which relief was obtained following hypnosis.

In both instances, the tinnitus was of the sort accompanying noise-induced hearing loss—middle to high frequency and bilateral—and the primary problem was interference with sleep. Different approaches were used with the two patients, but no explanation is given of the basis for choice.

One patient was instructed in a "mind over matter" approach while under hypnosis; he was instructed to concentrate on the tinnitus and note that its magnitude was diminishing. Posthypnotic suggestion was given that the patient could do this himself at bedtime with the same effect. Over a 6-week span, at least, the patient did find the tinnitus to be less of a problem, both at bedtime and during the day.

The second patient was told, under hypnosis, to recall a particularly pleasureable musical passage whenever his tinnitus became troublesome. The stated intent of this was to create an internally generated masking sound for the tinnitus, and, while it appears somewhat bizarre, the procedure apparently was very effective for this patient. During a 6-month follow-up period, the patient reported no instances of tinnitus interfering with his sleep—although he did on occasion notice some "strange music" at bedtime!

It is certainly believable that hypnosis could serve as an additional form of assurance to a patient or could produce a feeling of control over the symptom, thus supplying a sense of relief. It seems to be worth trying on some patients, for example, prior to beginning a program of psychotherapy.


It probably comes as no surprise that acupuncture has been attempted as a treatment for tinnitus. Largely undocumented claims of success can be found, but at least one author believes that only about 5 percent or less of patients are helped by acupuncture (Mann, 1974). Given that such values are probably in the range of spontaneous remissions and given that higher success rates are achieved with other procedures, there appears to be little justification for resorting to acupuncture for the treatment of tinnitus.

Electrical Stimulation

Hatton et al. (1960) found that 15 of an initial sample of 33 patients suffering from tinnitus reported increases or decreases in its magnitude under electrical stimulation applied through electrodes on the head and body. The positions of the stimulating and reference electrodes were systematically rotated in such a way that each subject was tested in eight conditions (two polarity arrangements for each of four pairs of electrode positions).

Consider first those patients, of the subset of 15, having monaural tinnitus. When Hatton et al. located ''the anode" on the same side of the patient's body as the tinnitus, the magnitude of the tinnitus was decreased, sometimes to the point of total suppression, as current strength was increased. With the electrode polarity reversed, these patients reported an increase in tinnitus magnitude as the current was increased. (Residual inhibition was not tested for or spontaneously noted.) In the patients with bilateral tinnitus, the same pattern obtained— the tinnitus on the side of the body with the anode" decreased and that on the side with "the cathode" increased as current strength was increased.

The patients whose hearing loss Hatton et al. traced to noise trauma and to presbycusis were notably unresponsive to the electrical stimulation. Apparently neither increases nor decreases in tinnitus were observed by these patients. If confirmed, the outcome will be an important one, for it has strong implications about the site of action of the electrical stimulation.

Tonndorf (CIBA Foundation, 1981:226) reported that J. D. Hood observed a suppression of tinnitus during and following electrical stimulation to the mastoids. Bilger (1977:169) reported that some patients fitted with an auditory prosthesis (cochlear implant) experience alteration of their tinnitus when the stimulating device is turned on; for some of these people the tinnitus is exacerbated, for others it is reduced or eliminated. For some patients these effects diminished with time since implantation. Brackmann (1981b) generally confirms Bilger's findings and implies that many implant patients experience pronounced residual inhibition following use of their stimulators. Brackmann also asserts that stimulation from the implant can reduce or eliminate contralateral as well as ipsilateral tinnitus and that the implantation operation itself reduced or eliminated preexisting tinnitus in about 80 percent of a particular sample of 29 patients.

Cazals et al. (1978) have confirmed earlier reports by others (e.g., Graham and Hazell, 1977; Field, 1893) that tinnitus can sometimes be abolished through the transtympanic application of brief electrical pulses to the cochlea (also see Portmann et al., 1979). Cazals et al. dealt only with patients having severe to profound hearing loss. The reference electrode was a silver disk attached to the earlobe; the active electrode was a ball of silver attached to the round window or promontory (although Aran and Cazals [1981] now urge the use of a platinum-iridium wire). The stimuli were electrical pulses of variable intensity, duration, frequency, and polarity. When "negative current" was applied to the active electrode, 81 percent (13/16) of the subjects were able to perceive auditory sensations of some sort with some combination of stimulus values; this effect has previously been reported and studied, of course, and is the basis on which hope for a cochlear implant prosthesis is built (e.g., Merzenich, 1975).

When the polarity of the electrical pulses was positive at the active electrode, auditory sensations were rarely reported, and, when they were, they were brief and they occurred at pulse offset—i.e., during a negative-going portion of the stimulus. However, in the subset of six patients who experienced tinnitus, five reported that it was suppressed by positive pulses. (Indeed, one of these subjects was not aware that he normally had a continuous tinnitus until it was eliminated by positive pulses.) As might be expected, the intensity necessary to achieve suppression varied across subjects (by a factor greater than 10); not so expected, perhaps, is the fact that frequency of stimulation appeared not to be important as long as it was higher than some "critical" value that did vary across subjects. The one subject not experiencing tinnitus suppression from positive pulses had had cranial trauma with pontine concussion, and thus it is possible that his tinnitus originated from a central location not affected by this electrode placement. An interesting report from one of the subjects who did have some residual hearing was that his understanding of speech was not obviously degraded by simultaneous presentation of the positive electrical pulses.

In a later report, Aran and Cazals (1981) make some additional points. The electrical stimulation was only effective in suppressing tinnitus clearly localized in the ipsilateral ear—never in the contralateral ear (compare Hatton et al., 1960). While no constant clinical features could be identified as predictive of the effectiveness of electrical stimulation, Aran and Cazals believe that the suppressible instances had a peripheral origin and the nonsuppressible instances had a central origin. In some patients the electrical stimulation was maintained for as long as an hour, with no diminution in effectiveness. Apparently some patients did report periods of relief extending beyond the duration of the electrical stimulation (so-called residual inhibition), but Aran and Cazals are clearly dubious about the reality of this effect. (Graham and Hazell [1977] once observed a 4-hour-long suppression following a 30-second stimulation.)

Cazals et al. (1978) suggested that the electrical stimulation with positive pulses produces the equivalent of hyperpolarization in neural elements of the cochlea or beyond. The idea is obviously in accord with the fact that reversing the polarity of the pulses—thereby presumably causing the equivalent of depolarization—produced auditory sensations in the same subjects. It does not explain why pulses are effective and steady-state (DC) stimulation apparently is not nor why there is a ''critical" frequency of pulse stimulation. A test of the idea with animals might be informative.

Thus, electrical stimulation appears to be worthy of further investigation as a treatment for severe tinnitus. The next step might be to try the procedure of Cazals et al. (1978) on volunteer tinnitus sufferers who have less hearing loss than did those patients. The reports from the Aran group raise the intriguing possibility of chronic implantation of both an electrode on the round window and a stimulator delivering pulses of the correct polarity for patients with otherwise intractable tinnitus. As Aran (1981) has noted, suppression of tinnitus with electrical stimulation may eventually prove to have diagnostic value even if it does not prove useful itself as a treatment.

A related matter deserves note here. Among mental health professionals there exist anecdotes about electroconvulsive therapy (ECT) producing tinnitus, sometimes in a severe form. No published reports could be found on this topic, but the relationship was recently noted (CIBA Foundation, 1981:230).

Alteration in Air Pressure

Vernon et al. (1980) briefly describe a device developed by R. Sandlin that has produced short-term relief in a single patient with a roaring, low-frequency tinnitus and a history of otosclerosis. Apparently the device fits tightly into the ear canal and is equipped with a syringe and a valve that allow the patient to increase the resting air pressure in the space between the device and the tympanic membrane. The procedure is described as awkward but very effective for this patient. It is asserted that it is fairly common to observe relief from tinnitus as a consequence of a change in the resting air pressure, as occurs in routine tympanometry (see, for example, Johnson and Goodwin, 1981).

Tjernstrom (1977) manipulated the relative air pressure in the middle and outer ears of patients with Meniere's Disease and found that an overpressure in the middle ear produced decreases in the tinnitus, vertigo, and low-frequency hearing loss that are characteristic of this disease, as well as decreases in the feelings of nausea and fullness. It is reported that attempts to incorporate such pressure changes into a treatment regimen for Meniere's Disease is being attempted in Tjernstrom's laboratory in Malmo, Sweden (J. Tonndorf, personal communication).

Wilson (1980a: Figure 12) has reported being able to reliably produce a low-frequency tonal tinnitus in himself under conditions of body tilt. His explanation involves the changes in acoustic impedance offered by the middle ear (stapes) to the cochlea that alterations in body orientation would produce. In normal body orientation, the impedance of the middle-/inner-ear boundary can be thought of as mass dominated, but, as body orientation is inverted and hydrostatic pressure altered, the impedance comes to be stiffness dominated. Wilson notes that the polarity of a reflection off a stiffness-dominated termination is opposite to that of a mass-dominated termination, and he argues that his low-frequency tinnitus becomes audible under body inversion because the output of an existing "active process" in his cochlea is driven into oscillation by the stiffness-terminated reflections. By Wilson's argument, his orientation-induced tinnitus ought to have an objectively detectable counterpart (a spontaneous otoacoustic emission; see "The Objective/Subjective Issue" in Chapter 2) that does not exist under normal orientation, but he has so far been unable to isolate it.

Wilson and Sutton (1981) also introduced pressure changes directly into the outer ears of a number of people having spontaneous emissions. Both increases and decreases in pressure generally produced upward shifts in frequency of the emission that were experienced (by the subjects whose emission was audible) as increases in pitch and as alterations in the tonal character of their tinnitus. The sound-pressure levels of the emissions could either increase or decrease with pressure change, but no concomitant changes in tinnitus magnitude were reported. Apparently no actual psychophysical measurements of tinnitus were made in these conditions.

Evans et al. (1981) worked with a guinea pig that had a spontaneous otoacoustic emission. They were also able to shift the frequency of the emission upward by introducing small (0.5 percent) increases or decreases in outer-ear pressure. Unlike the Wilson and Sutton (1981) demonstration, however, only reductions in the level of the emission were observed under conditions of altered pressure. Of course, we cannot know from this whether there were concomitant changes in subjective experience.

Thus, we have one brief report from Vernon et al. (1980) of a patient whose tinnitus can be reduced by altering the air pressure in the outer-ear canal, and, on the other side, we have Wilson and Sutton's (1981) report of no obvious perceptual change in tinnitus magnitude following air-pressure manipulations that produce increases and decreases in spontaneous otoacoustic emissions. Ignorance of the cause of the tinnitus in the Vernon et al. patient permits us to view these outcomes as not necessarily contradictory. That and the recurrence of anecdote about air-pressure changes and tinnitus reduction imply that mechanical attempts to relieve tinnitus may deserve greater attention for purposes of both diagnosis and treatment.

Tinnitus Maskers/Instruments

The idea of using an external sound to mask an obtrusive tinnitus dates back at least to Jones and Knudsen (1928), although Saltzman and Ersner (1947) are typically given credit for first implementing the idea. To a nonsufferer, there may seem to be something inconsistent in a sufferer's willingness to substitute one continuous sound for another--a point to which we shall return—but its effectiveness for some tinnitus sufferers is not to be denied. Indeed, many sufferers independently discover similar self-treatment strategies, e.g., mistuning an FM receiver and using the interstation noise as a masker while sleeping, choosing an office with relatively high background noise instead of a quiet one, etc. But while some tinnitus sufferers and some hearing professionals have long been aware of the value of masking, it was not until relatively recently that there was any serious attempt to try to capitalize on masking as a general treatment for relief of tinnitus (Vernon, 1977).

Why substituting a masking noise for the tinnitus serves as effective relief for some patients has been explained in a number of ways. Some patients comment that their tonal or narrowband tinnitus is simply more annoying than the broadband masker they choose. An important feature for some patients is the knowledge that the external masker is "a real sound" and under their control if they choose to alter it in some way—clearly not a feature of the tinnitus. This issue of potential control as it relates to annoyance has been studied by Glass and Singer, 1972; see "Annoyance of the Tinnitus" in Chapter 3.

The observation made by Saltzman and Ersner (1947) was that a goodly number of patients experienced relief from their tinnitus when fitted with a hearing aid. Their explanation was that the now-amplified external sounds acted to "drown out" the tinnitus. This observation is, of course, in perfect accord with the self-administrations practiced by many tinnitus sufferers and noted above—the use of broadband maskers such as FM interstation noise, the preference for a relatively noisy working environment over a quiet one, etc.

Nearly 30 years after the Saltzman and Ersner report, Vernon (1977) took their observation an additional step and suggested not just amplifying ambient sounds to achieve masking, but deliberately generating sounds for that purpose. The idea apparently originated in part out of frustration in treating tinnitus sufferers having reasonably normal hearing over the standard audiometric frequencies. These patients often had tinnitus at such high frequencies that masking it by using hearing aids to amplify ambient sounds was either impossible—due to the bandwidth limitations of the aids—or impractical—due to the consequent amplification-induced distortion in the low-and mid-frequency regions in which hearing was normal. But while the procedure was apparently originally conceived for use with these special cases of tinnitus, its use was soon extended so that tinnitus masking has now been attempted on sufferers having tinnitus of a wide variety of types.

In order to simplify the process of tinnitus masking in the everyday lives of their patients, Vernon and his collaborators developed sound-generating circuitry that could be housed in a standard hearing-aid chassis. This device was called a tinnitus masker. In a later development, tinnitus masking circuitry was housed in the same case with a standard hearing aid. This combination was called a tinnitus instrument; it has proved to be a useful device because many tinnitus sufferers also have hearing loss that is responsive to amplification. It is estimated that 10,000 tinnitus maskers and instruments are now in use in the United States (Vernon and Meikle, 1981).

If we presume for a moment that tinnitus will behave like an external sound when it comes to masking (see "Some Ways Tinnitus Is Not Like an External Sound" in Chapter 3), then, ideally, the sound generated to mask a tonal or narrowband tinnitus would itself be reasonably narrowband, and it would be located as accurately as possible in the spectral region of the tinnitus. A masker with these characteristics would be maximally efficient in that the least intensity would be required, and, thus, the least interference with other external sounds would result. While this ideal is nearly achievable with modern technology, to date it has not been vigorously pursued. Instead, the masking sounds generated by all tinnitus maskers/instruments are rather broadband, with limited energy in the region above about 6 kHz (see Vernon et al., 1977). Thus, for high-frequency tinnitus, greater intensity will be required for masking—if masking can even be achieved--and consequently there will be greater interference with the perception of real-world sounds than would be true if the masker were narrower and spectrally better located. Broadband or spectrally complex tinnitus, of course, raises the same problem in a slightly different form: in order to mask the tinnitus, other real-world sounds may also be partially or wholly masked, and it becomes increasingly difficult to decide whether the symptom or its treatment is the less desirable.

A recent finding by Penner (personal communication) is relevant to this issue of the ideal bandwidth for a tinnitus masker. Recall that she obtained daily pitch matches over the course of a month from three sensorineural subjects and that the range of settings covered thousands of Hertz for every subject (see "Pitch Matching" in Chapter 3). If this result is an accurate reflection of a long-term variability in tinnitus pitch, it might be the case that a narrowband masker fixed in frequency would not be the optimal tinnitus masker for some subjects; a tunable narrowband masker or a masker broadband enough to cover the range of frequency variation might be preferable. This matter clearly deserves further experimental attention.

Efficacy of Tinnitus Maskers/Instruments

The early reports on the efficacy of tinnitus maskers were very optimistic (Vernon et al., 1977; Vernon, 1978b; Vernon and Schleuning, 1978), perhaps misleadingly so, and, as a consequence, maskers began to receive much attention from auditory specialists and the popular press (Galton, 1979). Unfortunately, later reports (Schleuning et al., 1980; Roeser and Price, 1980; Rose, 1980) have been somewhat less encouraging, and this seems to have produced a widespread feeling that tinnitus masking is of little or no value (e.g., Loavenbruck, 1980). As will be seen, the truth lies somewhere between the extremes of panacea and worthlessness.

The reader is cautioned at the outset that at present there exists only one substantial set of data on the effectiveness of tinnitus maskers/instruments—the data from the Oregon group. This means that, even after a careful analysis of the data, erroneous conclusions are possible, for unreported procedural details or aspects of the data-reduction process may be instrumental in producing an undetected bias in one direction or another. We have, through private correspondence with the Oregon group, attempted to verify that the details of the published accounts are complete and, where not, to ascertain those details. Nevertheless, the risk remains that the conclusions reached here are appropriate, given the available data, but incorrect due to procedural or analysis details unrecognized for their biasing influence. Only additional, large-scale studies will reveal whether the following assessment of the effectiveness of tinnitus maskers/instruments is correct.

A procedural note is also in order at this point. Any effort to evaluate a new treatment for a long-neglected malady is subject to bias in the direction of overestimating its effectiveness. This bias stems from the natural gratitude and hopeful expectations of the previously neglected patients, and it is surely magnified when the patients also make a financial investment in the treatment, as, for example, the purchase of a masker, instrument, or aid. Ideally, attempts to estimate effectiveness should be made only after enough time has passed to allow the patients to make objective judgments about effectiveness that are relatively free from this bias. Of course, ideal conditions cannot reasonably be expected in real-world research, so it becomes the reader's job to remain alert to the possible contribution of this bias in what follows.

The Early Reports

In a 1977 article, the establishment of a tinnitus clinic at the University of Oregon Medical School was announced, along with some preliminary results (Vernon et al., 1977). At that time, 80 tinnitus patients had been seen; masking as provided by hearing aids had been recommended to 48 (60 percent) of these, and tinnitus maskers had been recommended to 29 (36 percent). (Tinnitus instruments did not become available until late 1978.) Of the patients accepting the recommendations, about 68 percent of the hearing-aid users and about 76 percent of the masker users "experienced total relief from their tinnitus" when the device was worn. About two-thirds of this latter group were also reported to have residual inhibition of their tinnitus. The implication given in this report is that the average patient routinely experienced 35–40 minutes of residual inhibition following a day's use of the tinnitus masker; at the other extreme, one patient had once experienced 16 hours of relief, and another, 18 hours.

By 1978, 158 patients had been seen at the Oregon tinnitus clinic (Vernon and Schleuning, 1978), and a third category of recommended treatments had been added—so-called FM masking, which is simply the use of interstation radio noise as a background when working quietly or trying to sleep. By then, about 52 percent of them had been recommended hearing aids, about 24 percent tinnitus maskers, and about 6 percent FM masking. (Thus, the proportion apparently receiving no recommendation was up from 4 percent in 1977 to 18 percent in 1978—a dramatic change, given that the 80 patients of 1977 were presumably included in the 158 of 1978.) According to the 1978 report, the percentage of patients who "obtained complete relief of their tinnitus" was about 67 percent (28/42) of those who accepted the recommendation of a hearing aid, about 81 percent (26/32) of those who accepted the recommendation of the tinnitus masker, and about 67 percent (6/9) of those who accepted the recommendation of FM masking. Vernon and Schleuning (1978) calculated that overall about 72 percent of their patients had been relieved of their tinnitus by masking in one form or another. Unfortunately, this 1978 report contains no update on the proportion of patients showing residual inhibition, but two cases are related in which the patients' residual inhibition gradually increased with continued use of the tinnitus masker, until the tinnitus eventually disappeared completely, causing the patients to return the masking units. Both of these "cures" involved tinnitus of several years' duration.

Later in 1978 Vernon (1978a) reported some summary statistics on the 513 patients who had, by then, been seen in the Oregon tinnitus clinic. Unfortunately, not all of the percentages of interest here were presented, but two trends did persist: the percentage to whom hearing aids were recommended dropped further, to 48 percent, and the percentage to whom no recommendation was made rose further, to 23 percent. Also of interest is the fact that about 78 percent of all patients seen exhibited either partial (43 percent) or complete (35 percent) residual inhibition in Vernon's standard test for residual inhibition (discussed under "Residual Inhibition" in this chapter).

Thus, the early evidence indicated that effective relief from tinnitus can be obtained through the application of masking. It must be appreciated that the patients seen at the Oregon tinnitus clinic were, by and large, an extreme sample—people with long-standing or particularly troublesome tinnitus—a fact that contributed greatly to the significance attached to the positive results by auditory specialists and laymen alike. It is fair to say, however, that many hearing specialists remained unconvinced. In part this skepticism stemmed from a reaction to the way the data were being released and to certain actions of the American Tinnitus Association (see McFadden, 1982). Nevertheless, the evidence for the effectiveness of masking against tinnitus continued to accumulate.

The Later Reports

By 1980, the Oregon tinnitus clinic had seen about 1,000 patients having tinnitus as their primary complaint. In an attempt to evaluate the longterm effectiveness of the various masking schemes that had been prescribed, questionnaires were sent to about 750 of these patients, and the 598 (80 percent) that were returned were discussed by Schleuning et al. (1980). For a number of justifiable reasons, these authors chose to discuss separately the data from patients seen from 1976 to 1978 and those seen in early 1979; these two subsets consisted of 493 and 105 returned questionnaires, respectively. Because it was based on a full-fledged questionnaire, this 1980 report is obviously more detailed than were the previous ones, and, while this detail is welcome, it does make comparison with the early reports difficult in places. Obviously, neither Schleuning et al. nor we can know whether, or how, the data would change if all the questionnaires had been returned. Summaries of the data reported by Schleuning et al. and discussed below can be found in Tables 1 and 2.

TABLE 1. Results of Follow-up Questionnaire Returned by 493 Patients Seen 1976–1978.


Results of Follow-up Questionnaire Returned by 493 Patients Seen 1976–1978.

TABLE 2. Results of Follow-up Questionnaire Returned by 105 Patients Seen in Early 1979.


Results of Follow-up Questionnaire Returned by 105 Patients Seen in Early 1979.

Of the 493 patients from the 1976–1978 period who returned the questionnaires, only about 27 percent (132) had been recommended hearing aids; this percentage is down from about 52 percent of the 158 patients described in the Vernon and Schleuning (1978) report and down from about 48 percent in the Vernon (1978) report, a discrepancy that is not discussed. (Specifically, in the Vernon [1978] report on 513 patients, 246 [48 percent] had been recommended hearing aids, yet in the Schleuning et al. report of the 493 patients in the 1976–1978 group, only 132 [29 percent] had been recommended hearing aids. Vernon [personal communications] could not account for this fall in absolute numbers, except to note that the two samples were not perfectly overlapping. The impression given in the various Oregon reports is that they are a series of updates and that the statistics are cumulative. Apparently this is not strictly true.) Approximately 41 percent of the 493 patients had been recommended tinnitus maskers, and an additional 9 percent had been recommended the newly developed tinnitus instruments; this total of 50 percent is more than double the 24 percent cited by Vernon and Schleuning (1978). Finally, about 23 percent were given no recommendation—very similar to the proportions reported earlier—but it is unclear whether or not this category now contains the patients who were recommended FM masking, a category used in the earlier report by Vernon and Schleuning (1978).

It is possible to estimate the success of these various devices in a number of ways. One is by examining the fraction of patients still using the device after a fixed period of time. Of the 216 patients in the 1976–1978 subset who did purchase the device recommended, 73 percent (158) claimed to be still using it at the time of the questionnaire. Another way to estimate success is the one used in the early reports—simply asking the patients whether or not the device provided relief from their tinnitus. In the early reports, the only positive outcome category was ''complete relief''; on the questionnaire, however, only 6 percent of the 1976–1978 subset reported "total" relief, with an additional 60 percent reporting "partial" relief (these percentages are calculated across the three devices—aids, maskers, and instruments—but they are much the same for each). Schleuning et al. noted that the interpretation of this outcome is unclear; it may have been that some patients had complete relief from tinnitus during the wearing of their devices (so-called active relief) but that they had acquired high expectations of pronounced residual inhibition that were not realized, causing them to regard their overall relief as only "partial."

The percentages cited above were only for the 1976–1978 subset of respondents to the questionnaire; the data for the 105 patients in the 1979 subset are different in a number of ways (compare Tables 1 and 2). For one thing, more tinnitus instruments and fewer hearing aids were recommended—about 24 percent and 16 percent, respectively, as compared with about 9 percent and 27 percent in the 1976–1978 subset. The explanation of this shift is straightforward—the tinnitus instrument was not commercially available until late in the 1976–78 period, and, since most of the patients did have some hearing loss accompanying their tinnitus, it just made good sense to recommend the combined aid/masker units once they became available. About 34 percent of the 1979 subset had only tinnitus maskers recommended to them, down from about 41 percent in the 1976–1978 subset. No recommendation was made for about 26 percent of the 1979 subset, up from about 23 percent in the 1976–1978 period. On the two measures of success, the 1979 subset surpassed the 1976–1978 group. Overall, about 92 percent of the patients who did purchase the recommended device claimed to be still using it at the time of the questionnaire—compared with 73 percent in the 1976–1978 subset—and about 15 percent reported "total" and 68 percent "partial" relief—compared with 6 percent and 60 percent, respectively.

Schleuning et al. (1980) cited a number of factors contributing to the improvement in the success figures from 1976–1978 to 1979, but they seemed to regard technological improvements in the tinnitus instruments as the primary one. Of course, an important variable operating to confound unambiguous interpretation of the apparent improvement is the time factor; at the time of the questionnaire, the 1976–1978 group had had more time than the 1979 group in which to become disenchanted with their masking devices. A possibly overriding factor has surfaced, however, that deserves special attention. Not mentioned in the Schleuning et al. article is a change in the Oregon group's procedures for making recommendations that was apparently instituted in late 1978 or early 1979 and thus differentially affected the success statistics of the 1979 group (Vernon, personal communication). The change was to make only a "tentative" recommendation after the initial examination. The patient took that initial recommendation to a hearing-aid dispenser and tried the recommended device for a month before deciding to purchase or return the unit. An unspecified number of patients returned to the clinic with complaints during this 1-month period, and some received new "tentative" recommendations. Apparently, a recommendation only became "final'' once a patient had accepted or rejected the "tentative" device and/or had ceased returning to the clinic. Thus, with the new procedure patients who would presumably have been categorized as unsuccessful under the 1976–1978 procedures were, in some cases, being reexamined, given alternative recommendations, and eventually counted as successful. No one interested in good care for the patient can dispute the virtues of this flexible, staged procedure for making recommendations; it is unquestionably a laudable change. But the change does create a statistical problem for those interested in evaluating the efficacy of masking for tinnitus. Specifically, the improvement seen in the success measures obtained from Tables 1 and 2 must now be viewed with some reservation. Given the modified procedures, it would be astonishing if the success figures did not improve from 1976–1978 to 1979—in the latter group the final recommendations sometimes came only after the failure of one or more tentative recommendations, and presumably some very risky cases were given tentative recommendations but still appear in the no recommendation category in Table 2. The reader is encouraged to keep this important procedural change in mind while reading the remainder of this section.

Since they are currently the primary source of information on the effectiveness of masking on tinnitus, the questionnaire statistics of Schleuning et al. deserve a final summary. Questionnaires were sent to 750 patients who had been seen since the inception of the Oregon tinnitus clinic in 1976. Replies were received from 598 of them (an 80 percent return rate). The authors argued that their procedures and instrumentation and the products available to their patients had changed so much by the beginning of 1979 that the survey data from the patients seen since then should be analyzed separately from those of the patients treated during the initial 3 years of the clinic. When the data are partitioned in this way, they support the contention that the tinnitus clinic has become more effective in recommending and/or fitting devices capable of alleviating tinnitus. Looking across the three types of device recommended—maskers, aids, and instruments—about 42 percent of the 1976–1978 subset and about 63 percent of the 1979 subset were using the recommended device at the time of the survey (see Tables 1 and 2); these numbers become 73 percent and 92 percent, respectively, when the denominator of the fraction is changed from simply the number of patients given a recommendation to the number of patients who actually purchased the respective device after the trial period. This improvement in use holds within, as well as across, the three categories of recommended device. Also indicative of improved procedures and recommendations from the 1976–1978 period to the 1979 period are the percentages of patients reporting either partial or total relief from their tinnitus. In the 1976–1978 subset, this group was 66 percent of the patients who actually purchased the recommended device (about 38 percent of the total to whom recommendations were made), and in the 1979 subset it was 83 percent (about 56 percent of the total receiving recommendations).

All of these indications of procedural improvements must be interpreted with an eye to the fact that the percentage of patients to whom no recommendation was made went up slightly, from 21 percent in 1976–1978 to 26 percent in 1979; possibly some of the diagnostically more risky patients who were given recommendations in the early years of the clinic were not later on, and this would naturally contribute somewhat to the higher success rates. Even though the sample size is still quite small, note should be taken of the greater apparent success of tinnitus instruments than of tinnitus maskers (Table 2), a trend that is worth watching. A final indication of improved procedures and success can be extracted from the Schleuning et al. data. In the 1976–1978 subset, about 22 percent of the patients for whom a device was recommended simply did nothing—not even use the device during a trial period (see Table 1). This fraction dropped to about 11 percent in the 1979 subset (Table 2). Greater persuasiveness may be the primary factor in this change, but it is certainly also in accord with a presumption of improved procedures or products. Vernon (personal communication) also credits this improvement partly to the increasing willingness of medical insurance companies to pay for tinnitus maskers/instruments.

From these questionnaire data, then, there emerges a picture that, if not as optimistic as the early reports would have led us to expect, is still encouraging. Masking does contribute some relief to about one-half (56 percent) of the patients for whom it is recommended. This is not the three-fourths (72 percent) claimed at one point (Vernon and Schleuning, 1978), but it is still significant.

The right-most column in Tables 1 and 2 is presented in an attempt to give perspective to the overall value of masking as a treatment for tinnitus. For these calculations, the denominator used was the total number of questionnaires returned; that is, it is an attempt to take into account the fact that no recommendation of masking is given to many patients following examination. Thus, it can be seen that the typical patient appearing at the doors of the Oregon clinic has about a 42 percent (8 percent + 34 percent) chance of realizing either partial or total relief from his or her tinnitus. Given the apparent severity of the symptoms in the patients referred to the clinic, this is a laudable success rate.

Other investigators of tinnitus masking have reported lower success rates than the ones of the Oregon group, making those reports important to examine before reaching a conclusion about the efficacy of masking as a treatment for tinnitus.

Other Reports

The article by Roeser and Price (1980) is frequently cited as containing evidence on the efficacy of tinnitus maskers that is contrary to that of the Oregon group. As shall be seen, this conclusion is open to question. Roeser and Price studied 52 patients referred to them because of serious tinnitus. Following audiological examination, one of three recommendations was made: a trial period with a tinnitus masker (65 percent of the patients), a trial period with a hearing aid (21 percent), or no trial period with either device (13 percent). (These percentages are revisions of those given by Roeser and Price, because they chose to include in the no recommendation category five patients who were in fact candidates for a masker unit, but who chose not to participate in a trial period after hearing a masker and judging it to be less tolerable than their tinnitus; including these cases in the no recommendation category compromises the concept of "recommendation" when the intent is to evaluate efficacy.) Unfortunately, tinnitus instruments were not included in this evaluation. Of the 23 patients who did use a tinnitus masker on a trial basis, 20 returned a questionnaire sent them at some time following the trial period (an 87 percent return rate). Roeser and Price report data on three items from their questionnaire, but unfortunately only for the 20 patients encouraged to try tinnitus maskers; no information is supplied on the 11 patients who tried hearing aids. Fifteen of the 20 respondents (75 percent) answered yes to "Does the masking unit totally cover your ringing?"; 6 of 20 (30 percent) answered yes to "Do you find the noise that the masking unit produces more tolerable than your ringing?"; and 5 of 20 (25 percent) answered yes to "Is your ringing reduced or absent when the masking unit is removed?" The last question is obviously aimed at the issue of residual inhibition and is easily interpreted. However, when reaching a conclusion about active relief, Roeser and Price choose to emphasize the second question rather than the first. Thus, they conclude that 30 percent (6/20) of the respondents are receiving active relief. This interpretation is unclear; the question most directly related to active relief would seem to be whether or not the tinnitus was ''covered" by the masker, and to this question, 75 percent of the respondents answered affirmatively. The matter might have been made less ambiguous had Roeser and Price asked a version of the question asked by Schleuning et al (1980): ''Are you currently using the device as a source of relief from your tinnitus?" The authors do reveal that 55 percent (11/20) of their respondents did purchase a masker following the trial period.

Thus, the conclusion reached by Roeser and Price (1980)—that the tinnitus masker is much less effective than the Oregon group indicates—is not strongly supported by the data they present. Further, the Oregon group has criticized that study on a number of grounds, including failure to include tinnitus instruments in the evaluation and utilization of only a limited number of early models of tinnitus masker. The issue of the relative tolerability of masking noise and tinnitus—addressed by Roeser and Price's second question—is an interesting one and worthy of additional study, but adequate evaluation of the issue requires high confidence that the masking sound being used in the comparison is the optimal one.

Rose (1980) has published a brief note on his experience prescribing tinnitus maskers. Unfortunately, the presentation is too vague about important details (total sample size, severity of tinnitus, bases for recommendations, effectiveness of hearing aids, etc.) to be of much help in evaluating efficacy of masking on tinnitus, but some facts are worth noting. Of the 31 patients who participated in a 1-month trial period, 32 percent later purchased a masking unit; of these, about 38 percent were still using the device, regularly or irregularly, at the time of the follow-up interview. Just as with the Roeser and Price study, it is believable that these percentages would have been higher if more masker models or tinnitus instruments had been offered to the patients.

Pulec et al. (1978) found that about 7 percent of the 950 new patients seen by them in an 18-month period had tinnitus as a major complaint. Twenty-eight of these patients were given a thorough tinnitus evaluation, which revealed that 23 (82 percent) had tinnitus that was maskable. Several case studies are presented, but no follow-up statistics are presented to aid in evaluating long-term efficacy.

Miller (1981) indicates that a controlled study of tinnitus maskers/instruments is under way.

Finally, a member of this working group (A. Shulman) is now engaged in a long-term study of tinnitus maskers/ instruments. Several hundred tinnitus sufferers have been seen, but to date only preliminary results have been published (CIBA Foundation, 1981:257–258). Preliminary analyses indicate that about 73 percent of those for whom a device (masker, instrument, or aid) was recommended did purchase it after a trial period (compared with the 68 percent of Schleuning et al.; Table 2 above) and that only about 11 percent of all patients were given no recommendation (compared with 26 percent in Table 2), but the follow-up data on current use and reported relief had yet to be analyzed at the time of this writing.


So, while some investigators feel that effectiveness of masking on tinnitus has been overestimated by the Oregon group, we find that there are procedural and instrumentation features of these studies that could easily have led to an underestimate of effectiveness, rendering this "negative evidence" less than compelling. On the other hand, essentially all of the positive evidence does come from only one group at this time, a situation that makes any cautious scientist uneasy. Some inconsistencies have surfaced in the Oregon group's statistical summaries, and some of their procedural changes have surely contributed to enhanced success figures. If we accept their data, it appears that when reasonable care is taken in the choice and fitting of an aid, masker, or instrument, about 83 percent (15 percent + 68 percent) of tinnitus sufferers who actually purchase the recommended device will appreciate partial or total relief from their symptoms (Schleuning et al., 1980; patients in 1979 subset). This corresponds to about 56 percent of those for whom a masking device of some sort is recommended and about 42 percent of all patients seen (see Table 2). Such percentages are clearly attractive given the state of affairs prior to the introduction of tinnitus maskers/ instruments. However, the reader is cautioned to draw only interim and conservative conclusions at this time, in recognition of the weaknesses in the currently available data. A conclusion that appears consonant with the available data, yet conservative, is that tinnitus maskers/instruments appear to offer hope of at least partial relief to a substantial fraction of tinnitus sufferers. Stronger conclusions must await additional data from both the Oregon group and others.

Residual Inhibition

Recall that one of Feldmann's (1971) many interesting observations about tinnitus and masking was that, following termination of an effective masker, the tinnitus often did not immediately reappear at its premasking magnitude; frequently several seconds of silence were realized prior to the return of the tinnitus (see "Related Masking Results" in Chapter 3). Feldmann appreciated the possibility this effect offered; he explicitly mentioned the prospect of possibly "training" the underlying mechanism so as to provide the tinnitus sufferer with periods of relief.

In the early reports from the Oregon group, this (now-named) residual inhibition was given much attention, and there seems to have been great hope among some members of the group that residual inhibition would frequently prove to be protracted or trainable and thus very valuable to a large number of tinnitus sufferers. Reports of patients who showed extended periods of residual inhibition—hours, days, and even "cures"—were repeated and picked up by the press (Galton, 1979). As a consequence, unrealistic expectations were possibly created among both tinnitus sufferers and hearing professionals. This surely was unintentional, but it is believable that some patients for whom masking was reasonably effective may have forsaken its use once it was realized that—contrary to expectations—prolonged periods of residual inhibition were not being produced. There is no doubt that very long periods of relief and even "cures" have been realized by a few patients following masker use, but they are rare (Vernon and Meikle, 1981). Fortunately, the later reports have tended to play down the more sensational instances of residual inhibition and instead have emphasized its value as a predictor of success of masking as a therapy, although the goal of possibly prolonging residual inhibition to the point of making it permanent is still mentioned (Vernon et al., 1980). The techniques that are being, or have been, tried to accomplish this end have not been specified (see Vernon and Meikle, 1981).

The experience of the Oregon group is that patients who show residual inhibition are more likely to profit from masking of some sort than are patients who do not show residual inhibition. A standard procedure to measure residual inhibition is obviously desirable, and the Oregon group claims to have developed one, although the details of its development are not clear from the published reports. First, masking of the tinnitus is attempted with tones and narrowband waveforms. Then these maskers are presented for 60 seconds at a level 10 dB higher than the level that was effective in masking the tinnitus. Upon termination of the masker, the patient apparently reports periodically on the magnitude of his tinnitus. The "usual" length of the residual inhibition period is said to be 25–45 seconds following a 60-second masking interval (Vernon et al., 1977), with about 35 percent of their patients showing complete residual inhibition for at least part of this time and an additional 43 percent showing only partial residual inhibition (Vernon et al., 1980). Exactly how this information is used for making recommendations is not clear. However, the absence of residual inhibition is not taken as definite evidence of lack of effectiveness of masking therapy (Vernon, 1981), and it does not invariably lead to the patient being placed in the no recommendation category (Vernon, personal communication).

Not only are statistics about the relationship between residual inhibition and recommendation category not available, but information about residual inhibition as a predictor of masker effectiveness is also lacking. For example, the latest report summarizing follow-up data on 598 patients of the Oregon clinic (Schleuning et al., 1980) does not present any correlations between residual inhibition as measured on the standardized procedure and later success with a tinnitus masker/instrument. It seems reasonable to expect that the magnitude and duration of residual inhibition will have at least some relation to the magnitude and duration of the masker producing it. The choices of 10 dB above masking level and 60 seconds duration are not explained in detail and may be worthy of further examination. Also of interest would be objective data on the return of tinnitus magnitude following masker termination—using (say) loudness matching of some sort instead of verbal reports.

Some aspects of residual inhibition deserve further study. It is commonly asserted that Feldmann (1971, 1981) frequently observed residual inhibition when the masker was presented to the contralateral ear, while the Oregon group never does (Vernon et al., 1977; Vernon and Schleuning, 1978; Shulman, 1981:212). The truth apparently lies somewhere in between. Vernon (personal communication) claims that Feldmann observes contralateral residual inhibition in only about 10 percent of his patients and that that fraction is in good accord with the experience of the Oregon group. Since the two groups use maskers of markedly different intensity and duration, this agreement is encouraging.

Vernon (CIBA Foundation, 1981:282; Vernon and Meikle, 1981) has noted that following a period of complete residual inhibition, some patients report that their tinnitus does not return gradually and monotonically, but rather it "bounces back" in intermittent spurts; this fact obviously has important implications for the underlying mechanisms.

Goodey (CIBA Foundation, 1981:288) has reported an effect analogous to residual inhibition following drug treatment of tinnitus. Upon withdrawal of a drug that has been effective against their tinnitus, about half the patients find that the tinnitus never returns to its previous level.

For the sake of completeness, the experience of Penner (personal communication) with residual inhibition should be noted. Recall that she works exclusively with patients who are categorized as having sensorineural hearing loss. In this select sample she finds that, following exposure to the masker, about one-third report no change, about one-third report decreased tinnitus (residual inhibition), and about one-third report increased tinnitus. Recently, the Oregon group also mentioned seeing increased tinnitus following exposure in a substantial fraction of their patients (Vernon and Meikle, 1981). The effect has been named "residual facilitation," but nothing about its diagnostic value has been presented.

Safety of Tinnitus Maskers/Instruments

The concern most frequently expressed about tinnitus masters and instruments is that their use may be producing hearing loss, or additional hearing loss. When considering the problem of hearing loss induced by exposure to intense sound, one must attend to several features of the stimulating waveform (see Kryter et al., 1966): its spectral makeup; its overall intensity and the distribution of intensity across the frequency regions present; the duration of exposure; whether the exposure sound is continuous or interrupted; and, if the latter, the nature of the intermittency. In all applications described to date, tinnitus maskers/instruments supply continuous sound, so the contribution of intermittency to exposure-induced hearing loss is a problem that can be ignored for the time being. This leaves the issues of the spectral characteristics, intensity, and duration of the exposure, about which the following generalization is relevant: for a noise with any given spectral characteristics, the greater the intensity and/or the greater the duration of daily exposure, the greater is the risk of an exposure-induced hearing loss. The question, then, is how great is the risk of wearing a tinnitus masker/ instrument?

Spectral Characteristics

Many long-known and well-documented facts of masking lead to the belief that the most efficient masker for a given sound—the one that will require the least intensity—will be one that is located in the same spectral region as, and has a bandwidth similar to (if a little wider than), that of the sound to be masked. It must be remembered that, for the most part, these basic data were collected on normal-hearing subjects and that Feldmann (1971), Penner (1980), Penner et al. (1981), and Vernon et al. (1980) have argued that at least some forms of tinnitus do not behave in all respects like "real" sounds when it comes to masking (see "Some Ways Tinnitus Is Not Like an External Sound" in Chapter 3). Nevertheless, explicitly stated goals of the masker/ instrument-fitting procedure are to determine and prescribe the masking waveform that best masks the tinnitus (Vernon et al., 1977; Vernon and Scheuning, 1978; Vernon and Meikle, 1981). One obvious practical reason for these goals is the belief that the narrower the bandwidth of the tinnitus masker, the less interference it will produce with the perception of important real-world sounds, such as speech. But to date, these goals have been partially frustrated by lack of availability of test equipment and of tinnitus maskers/instruments capable of generating waveforms with a wide variety of bandwidths and center frequencies. In this regard, tinnitus test equipment has improved more rapidly than have the tinnitus maskers/ instruments (Voroba, 1979b).

The spectral characteristics of several available tinnitus maskers were shown by Agnew (1979b) and Vernon and Meikle (1981); while they do offer somewhat different distributions of intensity across frequency, all are quite wideband in their outputs. This feature of present maskers/instruments is troublesome in a number of ways. Most obvious is that wide bandwidth maskers have more potential to interfere with the perception of important everyday sounds than do narrower maskers; this factor acts to hold down the intensity at which the user sets the masker/instrument and, thus, possibly diminishes its effectiveness as a tinnitus masker. Second, all things being equal, the more wideband the waveform is, the less efficient a masker it will be. This factor acts to increase the overall intensity at which the user sets the masker/instrument, for the less efficient the masker is—again, all things being equal—the greater its intensity will have to be for a fixed level of masking. And, the greater the intensity, the greater is the concern about exposure-induced hearing loss.

Two engineering aspects to the problem of individually tailoring a masking waveform to a particular patient deserve comment. One difficulty has been in the actual delivery of high-frequency energy to the eardrum (the high-frequency response of hearing aids has been similarly limited by the lack of availability of small microphones or speakers with efficient high-frequency responses). However, recent developments promise resolution of this aspect of the problem; Killion (1981), for example, has developed behind-the-ear and in-the-ear systems having a nominal 16-kHz bandwidth. Utilization of these systems in tinnitus maskers/instruments is to be encouraged.

The second aspect of the problem of individual tailoring of tinnitus maskers involves the process of obtaining the masker waveform. Until recently, the most common procedure for obtaining a narrow band of frequencies was to begin with a relatively wideband source and then filter out all frequencies but those in the spectral region of interest. Filtering schemes have evolved over the years, but it is still the case that a filter capable of supplying a reasonably narrowband waveform at a high center frequency is much too cumbersome to incorporate into an everyday device such as a masker/instrument. And even if this were possible, in order to permit true individual tailoring of maskers, the manufacturers would have to make available either a large number of devices with different masker center frequencies or else a wide array of (say) plug-in filter modules. An obvious solution to this problem is not to generate a wideband source and then filter it to obtain a narrow band, but to generate a narrow band in the first place. With digital synthesis and manipulation of waveforms, such a possibility exists, but manufacturers of tinnitus maskers/instruments have yet to take advantage of this approach. A small, easily wearable device could be produced that is programmable by the manufacturer or the dispenser to generate a waveform with the specific spectral characteristics determined to be necessary to produce the most efficient masker of that patient's tinnitus—that is, a "master" tinnitus masker/instrument. It is likely that the manufacturers have (rightly) been cautious about developing such devices because of uncertainty about the long-term efficacy of masking for relief of tinnitus. The results reviewed here suggest that such development may no longer be premature.

In regard to the issue of tailored narrowband maskers for tinnitus, the reader is reminded of the pitch-matching data of Penner and Voroba (see "Pitch Matching" in Chapter 3) implying that tinnitus frequency may vary greatly across time. If confirmed, this finding suggests that the ideal tinnitus masker/instrument might be either narrowband but user-adjustable in frequency, or else just wideband enough to "cover" the frequency excursions of the tinnitus of the particular patient.

Manufacturers contemplating development of new maskers/instruments that capitalize on digital synthesis procedures should be aware of recent evidence that indicts waveforms with very steep spectral skirts (McFadden and Pasanen, 1980; McFadden and Plattsmier, 1981a). Even if subsequent reports do not confirm that such waveforms are dangerous to the physiological mechanisms underlying frequency resolution, they should still be avoided because of their demonstrated ability to induce a short-term tinnitus (see "Possible Experimental Models of Tinnitus" in Chapter 2), whose cumulative and long-term characteristics are unknown. It would be ironic indeed if a new generation of tinnitus maskers/instruments were themselves responsible for inducing additional tinnitus. The evidence is still scanty, but it appears that narrowband waveforms having attenuation rates of about -70 to about -200 dB per octave at their "edges" are reasonably safe, but that slopes of -400 dB per octave (McFadden and Pasanen, 1980) and -600 dB per octave (Lummis and Guttman, 1972) should be avoided.

Intensity of Tinnitus Maskers/Instruments and Duration of Use

On the basis of survey and experimental research on the permanent hearing loss caused by sounds of various intensities, durations, etc., so-called damage/risk criteria (DRC) have been established (Kryter et al., 1966). For a fixed level of risk (proportion of population manifesting losses of particular magnitude at particular frequencies), the DRC specify the combination of sound intensity (measured on the "A-weighted" scale) and duration of exposure that must not be exceeded during an 8-hour work day in order to satisfy that risk criterion. For many occupational purposes, the U.S. government has designated 90 dBA and lower as "safe" for a working lifetime of daily exposures (Occupational Safety and Health Administration, 1974). If the continuous noise exposure is 95 dBA, the maximum allowable daily exposure is limited to 4 hours per day, and, if it is 100, 105, 110, or 115 dBA, the maximum allowable daily exposures are 2, 1, 0.5, and 0.25 hours, respectively. Ward et al. (1976), to take one example, seem to view these as acceptably safe values, but some investigators feel that a more conservative definition of "safe"—a "lower fence"—is necessary to adequately protect workers. For example, Passchier-Vermeer (1974) argues for 80 dBA for steady-state noises and daily 8-hour exposures (also see Burns and Robinson, 1970; Berger et al., 1978). Obviously, if the exposures are for periods greater than 8 hours or for more than 5 days per week, the specified intensities would have to be correspondingly decreased in order to maintain the same average loss of hearing. This is an important point, for apparently many users of tinnitus maskers/instruments wear them throughout their waking hours (Vernon and Meikle, 1981), and some even sleep with them on. (Contrary to general expectation, perhaps, is the fact that only about one-half of the people with serious tinnitus have difficulty falling asleep, and this seems to bear no simple relationship to the matched intensity of the tinnitus [see Vernon, 1977].) Thus, it is not possible to reach a decision about the safety of tinnitus maskers/instruments by simply comparing their maximum overall intensities with the standard damage/risk criteria—longer daily exposures and more successive days per week are involved.

A report issued by the U.S. Environmental Protection Agency (1974) includes procedures for translating the 8-hour occupational standards into corresponding levels for 24-hour exposures. As noted above, an important consideration is the intermittency of the exposure. Let us first consider an extreme, hypothetical case—a patient who uses a tinnitus masker/instrument continuously at the same level throughout each 24-hour period. For this no-intermittency situation, the translation procedure yields a value of 66.4 dB as the maximum permissible (in order to protect virtually every person from more than 5 dB of permanent hearing loss at 4000 Hz after 40 years of exposure). If the hypothetical patient did not keep the masker adjusted to the same level throughout the 24-hour period, but occasionally reduced its level or turned it off--that is, introduced intermittency—then the maximum permissible level could be greater than 66.4 dB. To take one example, assume that for at least 10 percent of each 1-hour period the patient sets the masker level to 65 dBA or less; the maximum permissible level now rises to 71.4 dB. The maximum permissible level rises to 73 dB if the masker is continuous for only 8 consecutive hours per day and is 60 dB or less for the remaining 16 hours. Various other examples are possible but are not necessary for current purposes. We see that somewhere in the vicinity of 66 to 73 dBA would be the maximum allowable level of a tinnitus masker that was worn regularly if existing damage/risk criteria were the relevant standard for regulation. It is to this range of levels, then, that one might want to compare the levels typically experienced by wearers of tinnitus maskers/instruments. Unfortunately, the comparison is not easily made, given the available data. (It must be emphasized that these EPA translations are viewed as unnecessarily conservative by many experts who believe that ''effective quiet'' may lie at or somewhere above 75 dB. Without getting into this controversy, it can be noted that, even if the safe level for continuous 24-hour exposure were to rise by, say, 15 dB from the 66–73-dBA range cited, none of the following arguments would be qualitatively altered.)

When dealing with pathological ears, the standard way to express sound level is relative to the patient's absolute sensitivity (absolute threshold) for that sound in that ear; sound levels so referenced are designated as decibels sensation level (dB SL). This usage is sensible for many purposes, for it often better communicates a sense of stimulus strength, or perceptual magnitude, in pathological ears than does a unit with a fixed reference such as decibels sound-pressure level (dB SPL). The problem this raises for us, however, is that DRC, effective quiet, etc., traditionally are specified in units analogous to dB SPL, while the intensities necessary to mask a tinnitus are typically given in dB SL without the necessary information to make the transformation to dB SPL. Thus, while we know from numerous reports that the measured loudness of tinnitus is typically low—only rarely is it matched to sounds greater than about 20 dB SL (see "Magnitude of the Tinnitus" in Chapter 3)—we do not know the overall sound-pressure level typically necessary to mask it. About all we presently have to go on when evaluating the safety of tinnitus maskers/instruments is the manufacturers' specifications of maximum level.

The manufacturers' specifications sheets for many models of tinnitus masker state a minimum overall output level of 40–45 dB SPL (note, not "A-weighted") and a maximum overall output level of between 85 and 95 dB SPL, and some models specify maxima in the range of 105–110 dB SPL. These measurements are sometimes made using a standard 2-cc coupler (e.g., HA-1), but they are frequently also made using a Zwislocki coupler (with the dB 110 adapter), which better reflects the high-frequency characteristics of the typical ear canal. This latter procedure is to be preferred, particularly when dealing with maskers/instruments specifically designed to mask high-frequency tinnitus, for the standard coupler underestimates the level of high frequencies.

Therefore, essentially every masker/instrument now available is capable of producing sound levels in excess of the 66–73 dBA that is derived as safe from existing standards, and some are capable of producing levels far in excess of that value. (Recall that current OSHA regulations would allow only about 0.5 hour per day per 5-day week of the 110 dB that some maskers/instruments claim to be capable of.) Thus, the potential for damage, or additional damage, to hearing clearly exists. At this time it is impossible to estimate how much risk is involved—or how much damage has already been done—because there are no systematic data yet available on the levels at which satisfied, chronic users of maskers/instruments routinely set their devices, on the temporal patterns of use, or on changes in hearing sensitivity after prolonged use of these devices. It may be that the typical user sets the masker/instrument to a level below the nominal 66–73 dBA or sets it to different values throughout the day, thereby reducing the overall risk. Data on these issues are important to acquire, and, given the large and growing number of users of maskers/instruments, they should not be difficult to obtain. People with bilateral hearing loss but monaural tinnitus would be particularly interesting, since the nontinnitus ear would serve as a within-subjects control for additional loss.

It should be noted that, if manufacturers do begin producing tinnitus maskers/instruments capable of generating narrowband waveforms, it will raise a further complication to the use of existing DRC and other guidelines as standards for regulating maskers/instruments. Most of what is known about exposure-induced hearing loss comes from exposures to relatively broadband waveforms, and it is that knowledge upon which DRC have been based. However, concentrating all the energy in a relatively narrow spectral region can create a potentially more damaging waveform than one with the same overall level but greater bandwidth. The magnitude of the additional risk is unknown, however. For such noises EPA guidelines in dBA are not appropriate, and criteria for octave or one-third octave bands such as those described by Kryter et al. (1966) should be used. To the extent that tinnitus behaves like an external sound when it comes to masking (cf. Feldmann, 1971; Vernon et al., 1980; Penner et al., 1981), the narrowband masker is to be preferred, for the overall level required to achieve masking of the tinnitus with a narrowband waveform should be lower than that required with a wide one—with external sounds it is the spectrum level that matters for masking, not the overall level. Even when the tinnitus does not act like an external signal, the narrowband masker has a virtue: it puts a smaller segment of the basilar membrane at risk of damage than does a wider-band masker. While there is some knowledge about the relative dangers of noise bands of different center frequency (Kryter et al., 1966), almost nothing is known about frequency regions above about 4000 Hz, so, if narrowband tinnitus maskers are developed for use at very high frequencies, existing DRC will be of questionable applicability.

Damage/Risk Criteria and Tinnitus Maskers/Instruments

Much of the above discussion may be irrelevant. It is extremely important to appreciate that compliance or noncompliance with current or future DRC may not be an appropriate basis on which to judge or to regulate tinnitus maskers/instruments. For many people, tinnitus is a severe, debilitating affliction, and it could easily be argued that relief from this affliction is, in many cases, worth the risk, or even the inevitability, of some hearing loss. Many commonly used drugs carry risks (or inevitabilities) of negative effects of various sorts, and even hearing aids themselves are capable of producing sound levels far in excess of both the 66–73-dBA and the 90-dBA values cited, depending upon the ambient levels in which the user lives. The implicit risks attendant to the use of such drugs, and of hearing aids, are generally judged by medical professionals, laymen, and users to be worth the benefits accrued. Tinnitus maskers/instruments appear to be indistinguishable from common drugs and hearing aids in this regard. That is, it can be argued that, if tinnitus sufferers are made aware of the potential risks to hearing involved in using a tinnitus masker/instrument at high intensities and/or for prolonged periods of time, and they still choose to wear the device, perhaps that decision should be theirs to make. It is clear that in the past "the consent" of masker/instrument users has not been as "informed" as it might have been, but this is true for users of hearing aids as well, and the solution appears simple. The practitioner, the dispenser, or the manufacturer could supply a brief written document that explains the situation and emphasizes the virtues of always using as weak a masker as possible and of introducing as much intermittency as possible (advice that seems equally desirable for users of hearing aids).

In this regard, the outcome of Penner et al. (1981) should be recalled; for some forms of tinnitus, continuous maskers lose their effectiveness relatively rapidly, and increasingly higher sound levels are necessary to accomplish the same result. Regular intermittency in the use of the tinnitus masker/instrument worn by such people would appear to be highly desirable. The ability to use lower masker levels ought to make the masker/instrument less obtrusive for the wearer, and it certainly ought to reduce the risk of masker-induced hearing loss. At this time no systematic studies have been done on the periodicity and duration of the intermittency necessary to control masker level in these patients.

Related to this matter of intermittency is another point. It might be possible to capitalize on the residual inhibition effect to gain greater safety and reduced inconvenience in tinnitus maskers/instruments. A member of the Working Group (L. Kaufman) realized that for some people it might be possible to periodically interrupt or to amplitude-modulate (AM) the masker wave-form at a rate that would allow residual inhibition to tide the person over until the next masker interval (or half-cycle of high intensity). The basic idea, of course, is to have the masker present only long enough and at adequate intensity to institute a subsequent period of residual inhibition. Interruption or AM rates of about two cycles per minute ought to be about right, according to the residual inhibition evidence reported by the Oregon group. The advantages to the user would be that: (1) about half the time the masker would be at a low level or absent and thus should be less of an impediment to the processing of real-world sounds such as speech, and (2) as a consequence, the average daily noise exposure would be reduced. The variants of the idea that come to mind appear to require more elaborate hardware than does simple AM, but obviously other temporal sequences of masker level would be possible with microprocessor-based maskers/instruments. Presumably, relatively long rise-decay times would always be important. A possible problem is that a slowly varying or regularly intermittent masker might be more annoying than a steady one, but the idea is definitely worth trying with some patients for whom residual inhibition is strong.

Hearing professionals have had a long-standing concern about the possibility that hearing aids induce additional hearing loss (sometimes known as induced deterioration or overamplification). Berry (1939) and Holmgren (1939) were among the first to raise the issue, which has yet to be satisfactorily resolved. To cite just a few of the published papers on the topic, Kinney (1961), Macrae and Farrant (1965), and Jerger and Lewis (1975) all concluded that there is evidence for temporary or permanent deterioration of hearing induced by hearing aids set for high amplification, while Naunton (1967), Bellefleur and Van Dyke (1968), Derbyshire (1976), and Titche et al. (1977) all concluded the opposite. A reading of this literature makes it believable that the disagreement stems in part from lack of adequate control of such important factors as actual level of amplification of the aid, length of its use, type of impairment, measurement procedures, and ambient environmental levels. In a recent article, Humes and Bess (1981) argue that—even using very conservative estimates of such critical factors as exposure level and duration—most hearing-aid users probably are at risk of substantial additional hearing loss from daily use of their aids. Their calculations are worthy of careful study by anyone interested in this important problem. For our purposes here, it must be concluded that, while the safety aspects of tinnitus maskers/instruments would appear to have much in common with those of hearing aids, too little is known about the latter to help us reach a decision about the former.

Copyright © National Academy of Sciences.
Bookshelf ID: NBK217858


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...