Clinical manifestations

Presentation and course

Clinical practice guidelines recommend that clinicians perform a targeted history and physical examination at the initial evaluation of a patient with subjective tinnitus (207). This will help to identify conditions that, if promptly identified and managed, may at least partially relieve the tinnitus (207). To prioritize interventions and facilitate discussions about natural history and follow-up care, clinicians should distinguish patients with bothersome tinnitus of recent onset from those with persistent symptoms for at least six months (207).

Key clinical features from the history include the patient’s age, onset (eg, insidious or sudden), duration, temporal pattern (eg, episodic or continuous, progression over time), location (ie, one ear, both ears, or nonlocalizable), pitch (high or low), amplitude (loud or soft), rhythm (steady, gradually fluctuating, or rhythmic), associated symptoms (eg, hearing loss, aural fullness, otalgia, vertigo, insomnia, anxiety, depression, headache, and neurologic dysfunction) as well as family history, previous head injury, noise exposure, medication use, previous ear infections, and previous ear surgery (34; 153; 177; 43; 101; 108; 91). Rhythmic tinnitus may be synchronous with the pulse or respiration, but although rhythmic tinnitus can be subjective, it is often objective. Tinnitus is generally most apparent and most bothersome at night when the masking ambient noise is less. Tinnitus severity is strongly related to tinnitus awareness during the day, self-reported depression and anxiety, subjectively experienced loudness, level of education, existence of additional somatic complaints, and subjectively experienced variability in loudness or pitch (81; 91). Tinnitus loudness is weakly correlated with pure tone averages at frequencies between 0.25 and 4 kHz, but much more strongly correlated with tinnitus annoyance and perceived impact on quality of life (01). Tinnitus self-reporting is inherently variable, and subjective tinnitus ratings fluctuate over time (32).

It is helpful if the patient can compare the sound of the tinnitus to an identifiable sound in the environment (101). High-pitched tinnitus may be described as ringing, steam- or wind-like, or clicking, whereas low-pitched tinnitus is often roaring, grinding, or like the sound of a seashell held to the ear. The tinnitus associated with Meniere syndrome is typically low-pitched (below 1000 Hz, and usually 125 to 500 Hz). When associated with middle-ear disease (including otosclerosis), it is typically low- or mid-range in frequency (250 to 2000 Hz); when associated with acoustic trauma or noise exposure, it is typically around 4000 Hz; when associated with presbycusis, acoustic neuroma, ototoxicity, and other sensorineural causes, it is typically high-pitched (2000 Hz and greater); and when associated with normal hearing, tinnitus of any frequency can occur (117; 33; 28; 119).

Tinnitus is frequently associated with depression (in up to half of patients) as well as anxiety, annoyance, anger, frustration, poor self-esteem, and insomnia (147; 44; 220; 21; 70; 97; 189; 134; 101; 91; 95). The causal direction is not always clear, as tinnitus may produce significant psychological stress, anxiety, and depression, whereas in some patients, tinnitus may result from depression or anxiety. Patients may excessively focus on their tinnitus, with detrimental effects on occupational and social functioning; these patients may rate their tinnitus as louder than nondistressed patients, even though such reports are inconsistent with objective measures of tinnitus loudness using masking (185). In general, older patients are more receptive to tinnitus, and most older patients with subjective tinnitus have concomitant hearing loss (152).

Patients with subjective tinnitus also report frequent sleep difficulties (84; 110). Compared to matched controls, those with subjective tinnitus show lower subjective sleep quality as measured by standard sleep quality indices and sleep diaries but no significant difference in objective polysomnograph sleep parameters (eg, sleep latency or efficiency) (69). Old age, daytime sleepiness, and psychological distress are strongly associated with tinnitus severity among patients with sleep difficulty (110).

Guidelines for grading tinnitus severity have been published (123), but no true consensus exists on the best approach to measure tinnitus (07).

Prognosis and complications

The prognosis depends heavily on the etiology. Most cases of subjective tinnitus cannot be cured, and many are refractory to treatment, particularly those that are bilateral or nonlocalizable. Patients may excessively focus on their tinnitus, resulting in detrimental effects on occupational and social functioning, but most individuals do not suffer marked activity limitation or participation restriction as a result of their tinnitus (168). Tinnitus is frequently associated with depression (in up to half of patients) as well as anxiety and insomnia. Although the causal direction is not always clear, tinnitus may produce significant psychological stress, anxiety, and depression. Chronic subjective tinnitus in older adults is subjectively louder, and more annoying and distressing, than chronic tinnitus in younger patients (04).

In a systematic review and meta-analysis of “no-intervention” periods in controlled trials, participants allocated to the no-intervention or waiting-list control arm of clinical trials for a tinnitus intervention show a small but significant improvement in self-reported measures of tinnitus with time (158). Whether this finding represents “regression toward the mean” or reflects some aspect of the natural history of subjective tinnitus is unclear.

Clinical vignette

Episodic unilateral subjective tinnitus with normal hearing: A 44-year-old woman presented with a 10-week history of episodic tinnitus in the left ear, which she described as a noise "like the screech of brakes of an underground train" (48). During attacks, she was unsteady, but she experienced no associated hearing loss, and between attacks, she felt normal. The attacks were brief (5 to 10 seconds) and occurred several times an hour without clear precipitants. Three years previously, she had experienced a period of unsteadiness for two weeks; one year previously, she had continuous left-sided tinnitus for 6 months. On examination during an episode, she was slightly unsteady but had no nystagmus, and her left corneal reflex and hearing were normal. Pure tone audiometry, speech reception thresholds, and tympanometry were normal. Brainstem auditory evoked responses were normal on the right but showed no reproducible responses on the left. Caloric responses showed a left canal paresis of 35%. Head MRI showed a left cerebellopontine angle mass that extended into the internal auditory meatus and compressed and indented the left pons. Treatment with carbamazepine abolished the episodic tinnitus and ataxia. The tumor was removed with a posterior fossa craniectomy. Histology showed a benign meningioma. Apart from slight ataxia, no significant neurologic residua presented, including deafness or further tinnitus, despite discontinuation of the carbamazepine.

Biological basis

Anatomic localization

Subjective tinnitus may occur with a wide variety of lesions of the external or middle ear, the cochlea or auditory nerve, or the central nervous system. It occurs most commonly due to cochlear dysfunction. Even with damage to the cochlea or auditory nerve, central auditory networks are disrupted, and the resulting abnormal central neural processes contribute to what is experienced as subjective tinnitus (03; 101). Evidence for structural abnormalities specifically related to tinnitus remains poor (03).

Pathophysiology

The pathophysiology of subjective tinnitus is poorly understood, but several mechanisms may be responsible (03; 101). With conductive hearing loss, internal auditory signals (eg, vascular noises and otoacoustic emissions) may be more apparent because of the ambient sound reduction. Cochlear damage (eg, from noise or ototoxins) or eighth nerve damage may result in abnormal afferent signals (eg, altered rate or rhythm), which are interpreted by the brain as tinnitus. Pressure on the eighth nerve may damage the myelin sheath, allowing ephaptic transmission or "crosstalk" between axons (48); usually, this mechanism is considered with posterior fossa tumors, but it has been argued that vascular anomalies (eg, a tortuous anterior or posterior inferior cerebellar artery) may also compress the eighth nerve and cause tinnitus (85; 86; 150; 26; 96; 116; 179; 180; 178; 213). Either absence or significant decrement in afferent auditory nerve impulses may also "release" central auditory network activity akin to visual hallucinations with blindness as well as "phantom limb" tactile and kinesthetic hallucinations following amputations (35; 176; 126; 62; 102; 11; 54; 27; 139; 218; 217; 183; 211; 31). Most adults experience tinnitus in anechoic environments (164; 40), perhaps because of such release mechanisms or perception of otoacoustic emissions from the cochlea that are normally masked by ambient noise.

Various evidence supports the concept that central auditory pathways participate maladaptively in the pathophysiology of tinnitus (101):

Subjective tinnitus may, thus, be caused by abnormal activity in a distributed network of peripheral and central pathways in the nervous system (03; 182; 64). The percept of tinnitus may be an emergent phenomenon of multiple partially overlapping neural subnetworks, each with a specific oscillatory pattern (which may have a pathologic component) and a characteristic pattern of connectivity (which may also be altered pathologically) (212). Abnormal neural activity in corticothalamic feedback loops or reduced inhibitory effectiveness between the limbic system and the thalamus are likely contributors to the abnormal central nervous system neural activity underlying subjective tinnitus (103). Central auditory and nonauditory brain regions are modified in tinnitus such that enhanced connectivity is evident in both auditory and nonauditory regions in the tinnitus brain, whereas controls without subjective tinnitus showed a decrease in connectivity following sound stimulation (182). A study using functional magnetic resonance imaging found that patients with subjective tinnitus had altered regional neural activity and inter-regional connectivity in partial auditory and non-auditory brain regions, mainly involving the “default mode” and audio-visual networks (64).

Studies using positron emission tomography and magnetoencephalography have suggested that tinnitus is related in part to plastic changes in the auditory association cortex (11; 111; 142; 54; 136; 27; 139; 184; 211). Tinnitus is associated with hyperactivity of the auditory cortices integrated in a global network of long-range cortical connectivity involving the prefrontal cortex, orbitofrontal cortex, and the parieto-occipital region (184); signals from the global network on the temporal areas correlate with the subjective strength of the tinnitus distress. There may also be a contribution from abnormal auditory-limbic interactions (188).

Severe tinnitus is almost always associated with hearing loss; tinnitus intensity levels established with loudness-matching techniques correlate with hearing levels at the frequency of most severe hearing loss (117; 149).

Even when hearing is apparently normal in patients with subjective tinnitus, there is usually associated cochleovestibular dysfunction. Abnormal otoacoustic emissions at higher frequencies suggest that there is mild (subclinical) outer hair cell dysfunction involving high-frequency cochlear regions (151). Unilateral canal paresis is also common in such patients, suggesting that tinnitus can be the only clinically apparent manifestation of a cochleovestibular lesion (141).

Epidemiology

One third of adults experience tinnitus during their lives; at least 10% of those develop prolonged tinnitus requiring medical evaluation (71), and in 0.5% to 2% of the population, tinnitus is disabling and significantly interferes with quality of life (125; 100). The point prevalence of non-pulsatile subjective tinnitus is approximately 10% to 15% of the adult population (30; 211; 100) and approximately 15% to 25% in older individuals (186; 55; 92; 133; 189). Episodic tinnitus is reported by approximately one fifth of patients in population surveys (66), but tinnitus generally tends to be persistent in more than 80% of affected individuals and progressive in some 40% over a 5-year period (55). The prevalence of chronic tinnitus in adolescents in the United States is about 5% (115). The incidence of tinnitus increases with age so that by the age of 70 years, at least 25% of patients experience constant tinnitus (186). Men are more often affected by subjective tinnitus than women, perhaps in part because of greater occupational and recreational noise exposure.

Reported risk factors for tinnitus include older age, male sex, symptomatic dizziness, hearing loss, exposure to ototoxic medications, hypertension, generalized anxiety disorder, a history of noise exposure, a history of arthritis, a history of head injury, a history of ever smoking or exposure to passive smoking, obesity, poor sleep quality, autonomic dysfunction, hypercholesterolemia, and low socioeconomic status (43; 56; 92; 133; 146; 189; 115; 101; 50; 120; 106). However, the reported risk factors are not uniformly consistent across studies (115). Even in the elderly, tinnitus may be associated with treatable health conditions (eg, otitis media, rhinosinusitis, ototoxic medications, depression), whereas other associations with tinnitus are less likely to be causal or to result in improvements in tinnitus symptoms if treated even if treatable (eg, hypertension, coronary artery disease, knee joint pain) (104; 133; 101). Although there is an association between tinnitus and hypertension, a cause-and-effect relationship is uncertain (50). In a systematic review, studies that analyzed the incidence of hypertension in tinnitus patients tended to show an association, whereas those that evaluated the incidence of tinnitus in hypertensive patients did not show an association (50). The reported association of poor sleep quality with subjective tinnitus (106) may indicate that subjective tinnitus disturbs sleep rather than poor sleep quality causes subjective tinnitus.

Subjective tinnitus may be an independent risk factor for cognitive impairment (165).

Genetic effects explain only a small proportion of tinnitus in a population (99).

The incidence rate of tinnitus in U.S. active-duty military service members rose consistently from 1.84 per 1000 in 2001 to 6.33 per 1000 in 2015 (140).

A very high rate of subjective tinnitus (46%) and of severe tinnitus (9%) was reported among Korean adolescents in a national survey coupled with otologic examination followed by pure tone audiometry up to 8 kHz (171). In this study, subjective tinnitus was associated with age, female gender, history of ear infection and sinusitis, recreational noise exposure (due to karaoke and local-area-network gaming), alcohol consumption, and cigarette smoking. Although hearing loss was not detected, participants with subjective tinnitus reported difficulty with sound localization and hearing in noise and suffered more physical and mental health problems than did those without tinnitus.

Differential diagnosis

Associated or underlying disorders

It is particularly helpful to determine if subjective tinnitus is unilateral or bilateral and nonlocalizable to determine if focal neurologic findings exist and to perform an audiogram to distinguish cases with conductive or sensorineural hearing loss (101). Subjective tinnitus may result from a wide variety of lesions of the external or middle ear, the cochlea or auditory nerve, or the central nervous system, but a large proportion are ultimately designated as "idiopathic" (105). The most commonly identified causal factor is noise-induced hearing loss. An important treatable cause of tinnitus is otosyphilis, which can occur in HIV-positive individuals despite high CD4 cell counts (137).

Tinnitus associated with conductive hearing loss (unilateral or bilateral) may be caused by impacted cerumen, osteomas, glomus jugulare tumor, otitis media, otosclerosis, or otological trauma (12; 58; 154; 172; 101). Glomus jugulare tumors are rare vascular tumors arising from paraganglia of the ninth and tenth cranial nerves. Patients with glomus jugulare tumors may complain of hearing loss, ear and facial pain, pulsatile tinnitus, dysphagia, and dysphonia; examination may demonstrate conductive hearing loss, lower cranial nerve palsies, and possibly a reddish retrotympanic mass. In rare cases, glomus jugulare tumors may secrete catecholamines, and patients may develop secondary hypertension, anxiety, headache, and glucose intolerance (206). Childhood otitis media with an associated hearing loss in the low and high frequencies is associated with tinnitus in adulthood (38).

Unilateral subjective tinnitus associated with progressive sensorineural hearing loss should suggest the possibility of an acoustic neuroma or other eighth nerve lesion (101). Other causes of subjective unilateral tinnitus associated with sensorineural hearing loss include Meniere syndrome, labyrinthine concussion, neurolabyrinthitis, autoimmune hearing loss, and perilymphatic fistula (72; 29; 101). Microvascular compression of the cochlear and vestibular nerves in the cerebellopontine angle or the internal auditory canal has also been considered a cause of various audiovestibular symptoms including fluctuant, pulsatile, or continuous tinnitus (usually unilateral) as well as hearing loss, hyperacusis, diplacusis, vertigo, and imbalance (85; 86; 26; 96; 179; 180; 178; 213), although the existence of this disorder remains controversial (02; 24; 116; 178). A specific clinical presentation for microvascular compression, if any, has not been clearly defined (02; 24; 116), although associated brief vertiginous spells and, particularly hemifacial spasm, have been considered by proponents to be suggestive features (179; 180). Acute-onset tinnitus with sudden deafness is often due to viral neurolabyrinthitis, labyrinthine ischemia, or labyrinthine concussion. Episodic subjective tinnitus may occur with Meniere syndrome, perilymphatic fistula, or lesions of the cerebellopontine angle (48; 200).

Bilateral subjective tinnitus associated with sensorineural hearing loss may be caused by ototoxic drugs, noise exposure, or presbycusis (135; 127; 159; 209; 28; 57; 177; 43; 189; 101). Commonly used drugs that cause tinnitus include caffeine, salicylates, nonsteroidal anti-inflammatory agents (eg, indomethacin), aminoglycosides, aminophylline, loop diuretics (eg, furosemide), quinidine, and platinum-based chemotherapeutic agents (eg, cisplatin and carboplatin) (43). Tinnitus is common with noise-induced hearing loss, which may be the most troubling symptom for some patients (117; 135; 216; 127; 159; 57).

Central nervous system causes of subjective tinnitus are often associated with central neurologic findings: head trauma, brainstem stroke, posterior fossa tumors (eg, acoustic neuroma, meningioma, metastatic tumors, and meningeal carcinomatosis), pineal region tumors, multiple sclerosis, vascular abnormalities (eg, arteriovenous malformations) or vascular tumors (eg, glomus jugulare), palatal myoclonus, meningitis, and pseudotumor cerebri (105; 41; 138; 48; 101).

Diagnostic workup

Clinical examination should include fundoscopy, otoscopy, tests of hearing, neurologic examination, auscultation for objective tinnitus (eg, from arterial bruits or venous hums), palpation of the neck or oral cavity for masses, and examination of the temporomandibular joint (101). Otoscopy can identify impacted cerumen, a perforated eardrum, middle ear fluid, cholesteatoma, and other masses. Auscultation is performed by listening with a regular or electronic stethoscope over the patient's external auditory canal, orbit, cranium, and neck. If objective tinnitus is identified, it should be timed relative to the pulse and respiration. A venous hum may be eliminated by turning the head or by gentle temporary pressure on the neck that occludes the jugular vein, but not the carotid artery. Blood studies can include a complete blood count, serum lipids, blood sugar, thyroid stimulating hormone, sedimentation rate, Lyme titers, and syphilis serologies (124; 34; 153; 67; 101).

Audiometry is essential in evaluating subjective tinnitus (101; 207). At a minimum, pure-tone audiogram should be performed along with assessment of speech reception thresholds and word recognition. The pattern of hearing loss, when present, is helpful in identifying possible causes. Tinnitus associated with conductive hearing loss may be caused by impacted cerumen, osteomas, otitis media, or otosclerosis (12; 58).

Several common configurations or patterns of sensorineural hearing loss exist that include: a (usually bilateral fairly symmetric) notched pattern (eg, noise-induced hearing loss), a symmetric bilateral downward sloping pattern (eg, presbycusis), and a low-frequency trough pattern (eg, Meniere syndrome).

In those with bilateral subjective tinnitus, noise-induced hearing loss and presbycusis are among the most commonly identified causes. Retrocochlear tinnitus should be suspected when sensorineural hearing loss is asymmetric or when word recognition is asymmetric or decreased out of proportion to threshold hearing level.

Tympanometry and acoustic reflexes are sometimes recommended (34), but often add relatively little to the evaluation of subjective tinnitus or retrocochlear hearing loss; they are less sensitive and specific than brainstem auditory evoked responses or MRI in the detection of acoustic neuromas (28).

Otoacoustic emissions can help differentiate between cochlear and retrocochlear sensorineural hearing loss (28). They generally reflect cochlear pathology (assuming no significant conduction abnormality) and are relatively insensitive to neurologic abnormalities. In particular, otoacoustic emissions can help interpret absent responses on brainstem auditory-evoked potential testing (28). If evoked responses are absent and otoacoustic emissions are normal, the problem is neurologic and cannot be attributed to peripheral hearing loss, whereas if both are absent, then hearing loss is probably peripheral.

Brainstem auditory evoked responses (BAERs) (or auditory brainstem responses [ABRs]) can be helpful in those with subjective unilateral tinnitus and either normal hearing or unilateral sensorineural hearing loss of unclear duration and course. Interpeak latency (IPL) III-V may be prolonged in tinnitus ears compared to non-tinnitus ears (63). Wave I amplitude may be smaller in tinnitus ears compared to non-tinnitus ears (63). An increased wave I to wave V interval (IPL I-V) should prompt MRI scanning to exclude acoustic neuroma.

A routine MRI is of little or no value in patients with tinnitus with persistent complaints, particularly with normal audiology or stable symmetric sensorineural hearing loss, and clinical guidelines recommend against ordering an MRI to evaluate tinnitus that does not localize to one ear, is nonpulsatile, and is not associated with focal neurologic abnormalities or an asymmetric hearing loss (207; 80). Note that anterior inferior cerebellar artery loops are often encountered on MRI scans but rarely cause tinnitus and should be considered incidental findings (80).

Patients with subjective tinnitus and either focal neurologic findings or progressive hearing loss should have an MRI with and without gadolinium contrast to exclude acoustic neuromas or other tumors, multiple sclerosis or other demyelinating disease, and brainstem stroke. A retrospective study suggested that MRI may help identify demyelinating disease among some adolescent children with sensorineural hearing loss, but this study is really inadequate to establish the utility of MRI in this patient group, particularly as it does not adequately address the existence of focal neurologic findings in cases with abnormal MRI findings (107).

Referral to an otolaryngologist should be considered for patients with tinnitus and either (1) conductive hearing loss (not attributable to impacted cerumen or otitis media), (2) mixed hearing loss, or (3) a retrotympanic mass.

Despite thorough evaluation, the cause of tinnitus is not identified in many patients, particularly in those with nonlocalized and continuous tinnitus or those with normal hearing, without neurologic findings.

Management

Prevention of hearing loss or tinnitus. Opportunities should be sought to prevent hearing loss and associated tinnitus (101; 181). In particular, no effective medical or surgical therapy is available for ototoxic or noise-induced hearing loss, which are the most common causes or contributors to subjective tinnitus prevalence; therefore, prevention is essential (101; 181). The patient's prescribed over-the-counter medications should be reviewed to identify possible sources of ototoxicity (eg, aspirin, aminoglycosides, anticancer drugs, ibuprofen, quinine, etc.) and substances that may exacerbate tinnitus (eg, nicotine, caffeine, alcohol, marijuana) (28; 20; 101). Subjects at risk for noise-induced hearing loss (eg, workers in heavy industry, heavy equipment operators, users of firearms, users of power tools, professional musicians) should use hearing protection (101). As many affected individuals are reluctant to wear hearing protection, education is critical; if feasible, environmental controls should be placed to control noise exposure (101). Hearing protection should be worn if individuals have to raise their voices to easily carry on a conversation. Hearing protection devices commonly available include ear muffs or ear plugs. Ear muffs usually have higher noise-reduction ratings than earplugs; they can be worn in combination in extremely high noise environments, but the maximum noise reduction is approximately 50 dB because of bone conduction through the skull. Misfit earplugs or ear muffs are much less effective and may not adequately protect hearing. Hearing aids (most of which are vented), cotton balls, and earplugs designed for swimming are all inadequate for hearing protection. Tinnitus, at the time of annual occupational audiometric testing, may help identify workers at increased risk for developing noise-induced hearing loss; it may occur before development of a 4000-Hz noise notch (57).

Symptomatic treatment of tinnitus. Despite a wide variety of treatment options, available therapies for subjective tinnitus generally have low success rates, leading to frustration by both patients and healthcare providers (61). In addition, available randomized trials of tinnitus interventions are most applicable to older adults with chronic persistent tinnitus who are evaluated in specialty settings (161). Various forms of bias, almost universally short follow-up periods, and problems with outcome reporting in available trials raise concerns about the validity and generalizability of the findings, which, in turn, may influence how clinicians apply trial results to individual patients and establish treatment expectations (161).

Patients reporting tinnitus have significantly worse hearing than those who do not report tinnitus. Two thirds of patients report tinnitus as annoying, yet only about one third seek professional help, and less than 10% receive treatment (193). Nevertheless, symptomatic treatment of the tinnitus, as well as treatment of any associated depression, anxiety, and insomnia, can be helpful even if the associated tinnitus cannot be eliminated (44).

Symptomatic treatment of tinnitus can include reassurance (195; 214; 15), various psychological treatments (eg, counseling, supportive psychotherapy, biofeedback, hypnosis, habituation) (36; 82; 121; 15; 100), masking techniques (185; 47; 06; 195; 79), electrical stimulation (197), drugs (195), and (rarely) surgery (195). The most widely employed treatments include reassurance, masking techniques, and various drugs, but many patients receive little apparent benefit from these (42; 160; 174; 79; 15).

Cognitive behavioral therapy. Clinical practice guidelines recommend that clinicians recommend cognitive behavioral therapy to patients with persistent, bothersome tinnitus, although available studies are generally of low methodological quality (207). There is limited evidence that cognitive behavioral therapy improves anxiety, health-related quality of life, or negatively biased interpretations of tinnitus (53). Cognitive behavioral therapy may reduce the negative impact that tinnitus can have on quality of life and may have a small additional benefit in reducing symptoms of depression, although uncertainty remains due to concerns about the quality of the evidence (53). Adverse effects appear to be rare in adults with tinnitus receiving cognitive behavioral therapy (53).

Psychologically based therapies have a variety of competing labels, with overlapping constellations of approaches that are hard to extricate into component treatments. Cognitive behavioral therapy, which uses relaxation, "cognitive restructuring of thoughts," and repetitive exposure to exacerbating situations to promote habituation to tinnitus, does not reduce perceived tinnitus loudness but may improve depression, anxiety, and “stress,” and thereby contribute to improved quality of life (121; 100; 109). Tinnitus retraining therapy, which uses "directive counseling and sound therapy in a strict framework," may be helpful based on the results of a single, low-quality controlled trial (157; 98). Tinnitus retraining or habituation therapy that utilizes a multidisciplinary treatment team shows limited promise but is expensive and not widely available (195; 22; 75; 76; 74; 68). A related technique of frequency-discrimination training was not helpful (78). Biofeedback does not reduce tinnitus loudness, but it may reduce muscular tension and anxiety, may facilitate coping with the tinnitus, and may be better accepted by patients than psychological counseling (06; 36; 153).

Augmenting hearing. Removal of ear wax may help some patients with tinnitus, partly because it increases ambient sound and assists with masking and because ear wax on the tympanic membrane can produce tinnitus through local effects on the conduction pathway.

According to clinical practice guidelines, clinicians should recommend a hearing aid evaluation for patients with persistent, bothersome tinnitus associated with documented hearing loss (207). A hearing aid may result in improved hearing and communication and may decrease the tinnitus by amplifying ambient sound and providing some masking (52; 143; 205; 187; 15; 98; 191; 73; 09).

Electrical stimulation with a cochlear implant or with transtympanic stimulation of the cochlea has been reported to be helpful in some patients, particularly those with unilateral severe to profound hearing loss (166; 93; 203; 23; 167; 202), and does not appear to exacerbate tinnitus in the contralateral ear, but high-quality studies are still lacking (166; 167). Transcutaneous electrical stimulation for subjective tinnitus does not appear to be helpful (90).

Extraneous sounds or noise. Various therapies based on the delivery of extraneous sounds or noise have been advocated (185; 52; 199; 98; 73; 18), and clinical practice guidelines indicate that clinicians may optionally recommend sound therapy to patients with persistent, bothersome tinnitus (207). For some patients who are particularly bothered by tinnitus at night, a bedside radio (possibly with a pillow speaker) tuned between stations can provide effective masking and allow them to get to sleep. White noise generators, white noise tapes, and tapes of "ocean surf" produce similar results (199). Some devices that provide a broad-frequency, customizable acoustic stimulus are available (37; 65). Further, tinnitus-masking devices that can be worn like a hearing aid are available as well as "tinnitus instruments" that combine a masking device and a hearing aid (52; 199; 98; 73). Masking is generally attempted ipsilaterally and contralaterally. Unfortunately, long-term efficacy of tinnitus masking units is at best variable and, in some studies, poor (185; 47; 34; 153; 79). In one large observational study of tinnitus masking, two thirds of patients achieved at least partial relief, but only 6% achieved complete relief (185), whereas results in one small, randomized trial were no better than placebo (47). In some patients, a high level of masking is required (ie, above 10 dB), and the patient finds the masking signal too distracting (28). In others, bilateral tinnitus is not relieved with masking devices, or the application of bilateral masking devices makes it difficult for patients to hear environmental sounds (28). In a systematic review, the limited available data assessing the efficacy of sound therapy (masking) in managing tinnitus in adults did not show strong evidence of efficacy, but no side effects or significant morbidity were reported (79). In addition to the traditional masking approach using white noise or unmodulated pure tones, modulated sounds can be tried for tinnitus suppression because amplitude-modulated tones with carrier frequencies near the tinnitus pitch may be more effective in reducing hyperactive neural activities associated with tinnitus (169). The availability of customized settings with sound generators may be helpful (18).

Pharmacologic therapy. Pharmacologic therapy for tinnitus has generally been disappointing: in general, drug therapies to date have not been successful for long-term tinnitus reduction (144; 28; 67; 15; 101; 207). Clinical practice guidelines recommend that clinicians not use antidepressants, anticonvulsants, anxiolytics, Ginkgo biloba, melatonin, zinc, or other dietary supplements for treating patients with persistent, bothersome tinnitus (207; 155). Similarly, no benefit has been demonstrated for antioxidant therapy in the treatment of subjective tinnitus (162).

Because intravenous lidocaine can provide transient relief of tinnitus (ie, usually less than 30 minutes) (129; 118; 83; 46; 156; 131; 196; 89), many investigators have evaluated a range of drugs in the hope that a long-acting oral medication could be found that was effective in the treatment of tinnitus (144). In particular, various anticonvulsants have been tried without much success (128), including gabapentin (19; 160; 17; 39; 201), carbamazepine (46), phenytoin, barbiturates, benzodiazepines (105; 144; 88; 214), and sodium valproate (130). Sulpiride, a D2 antagonist of dopamine receptors, in conjunction with either melatonin or hydroxyzine, significantly decreased tinnitus perception in short-term randomized, placebo-controlled, double-blind trials (112; 113). Other agents purported to be effective include nortriptyline (198), melatonin (1 to 3 mg per day) (175; 132), zinc (10; 148), acamprosate (a glutamatergic and GABA neurotransmission modulator used in alcoholism treatment) (14), and misoprostol (219). However, all of these agents provide, at best, modest relief in subsets of patients. Available evidence indicates that ginkgo biloba (77; 170), paroxetine (173), trazodone (42), memantine (51), and intratympanic steroid injections (08; 190; 204) are not effective. Clinical trials of pharmacologic agents have had significant deficiencies; methodologically improved studies are needed, using specific etiologic groups and appropriate controls with subjective and objective psychophysical measurements (144).

Clinical practice guidelines recommend that clinicians not use intratympanic medications to treat patients with persistent, bothersome tinnitus (207). Although intratympanic steroid injection has proponents (190), available studies have not unequivocally indicated any significant clinical benefit (08; 190; 204). In a prospective randomized, single-blind trial of medically refractory patients comparing intratympanic steroid injection (0.5 ml prednisolone or 0.5 ml of dexamethasone) with oral carbamazepine treatment, there were no statistical differences in the effective and control rates among the three groups (190). In another study of 70 adult patients with subjective tinnitus of cochlear origin who were randomly assigned to receive intratympanic injection of either methylprednisolone or saline solution, no significant posttreatment changes in the tinnitus severity index individual and total scores were observed in either group: thus, intratympanic methylprednisolone had no benefit, compared with placebo, for the treatment of subjective tinnitus of cochlear origin refractory to medical treatment (204). Moreover, the procedure is not without significant potential side effects, including pain during injection, vertigo, a burning sensation around the ear and in the throat, and a bitter taste (204).

Occasional patients may have specific types of subjective tinnitus that are likely amenable to pharmacologic therapy. Anecdotal reports suggest that low-dose carbamazepine, although not generally effective for subjective tinnitus, can relieve symptoms in patients with specific auditory hyperactivity disorders of the eighth nerve (eg, so-called "typewriter tinnitus" and "paroxysmal staccato tinnitus") possibly resulting from neurovascular compression akin to trigeminal neuralgia and hemifacial spasm (25; 145; 122). These cases are unusual but are worth identifying as a potentially effective pharmacological treatment is readily available (25; 145); complete long-term symptomatic relief has also been anecdotally reported following microvascular decompression of the vestibulocochlear nerve (122). One patient with basilar invagination had associated kinking of the auditory nerves at the porus acusticus with associated bilateral asynchronous clicking tinnitus resembling a typewriter (145). Four other cases reported frequent, repetitive, brief attacks, usually lasting only 10 to 20 seconds, of loud monaural tinnitus with a staccato character resembling clattering or machine-gun fire (25). Attacks occurred as often as once a minute and were generally spontaneous but could be precipitated by loud sounds or certain head positions (25). In patients with tinnitus and vitamin B12 deficiency, intramuscular vitamin B12 may produce improvement in perceived tinnitus severity (194).

Surgical treatments. Surgical treatment of subjective tinnitus has generally been disappointing. Nevertheless, a variety of surgical procedures have been advocated, including cochlear implantation (209; 210), cochlear nerve section (163), stapes surgery for otosclerosis (12), and surgical microvascular decompression (85; 86; 26; 96; 179; 180; 178; 213). Surgical case series have reported variable results, but many cases with subjective tinnitus do not benefit from surgery with a reduction in tinnitus, even if the surgery was necessary and helpful in other ways; some actually worsen (163; 209; 16). Indeed, tinnitus may increase in a substantial proportion of patients following acoustic neuroma surgery or stapedectomy for otosclerosis (72; 29; 05; 49). Even ablation of the cochlea or eighth nerve may not substantially alter the tinnitus in some patients with dysfunction of these structures (163; 16), suggesting that tinnitus may be maintained in central auditory pathways, even if it had its genesis in dysfunction of peripheral auditory structures. Patients most likely to develop postoperative tinnitus are those with poorer hearing preoperatively or surgically-induced hearing loss (29).

Although several authors have reported anecdotal success with surgical microvascular decompression for tinnitus (85; 86; 26; 96; 179; 180; 178; 213), the procedure remains controversial (02; 24; 116; 178; 59; 60). Radiologic demonstration of contact between a vascular loop of the anterior inferior cerebellar artery and the eighth cranial nerve on MRI is a normal anatomic finding and should not be used as the basis of diagnosing a "vascular compression syndrome" by itself (116; 60), although this still forms the basis for diagnosis of most reported cases (180). There is no significant relationship between subjective tinnitus and the type of vascular look, the vascular contact, or the angulation of the vestibulocochlear nerve at the cerebellopontine angle (60).

Treatment options for glomus jugulare tumors include surgical excision, radiotherapy, and embolization (206).

Other therapies. In a pilot trial, transcutaneous electrical nerve stimulation (TENS) applied to the auricula (of one or both ears) has a therapeutic effect on subjective chronic tinnitus as well as a placebo effect (208). A small (unblinded) randomized trial involving 46 patients (23 each in the treatment and control groups) reported that TENS is an effective, practical, and reliable therapy method for reducing the severity, loudness, and annoyance of tinnitus and for increasing the quality of life in individuals with chronic subjective tinnitus (13).

A small single-blinded trial of neurofeedback concluded that alpha NFB training can effectively reduce tinnitus-related distress and sound perception in patients (114).

Other not recommended therapies. Clinical practice guidelines recommend that clinicians not use transcranial magnetic stimulation for the routine treatment of patients with persistent, bothersome tinnitus (207). Nevertheless, studies continue to explore potential utility of transcranial magnetic stimulation for tinnitus relief (192; 215).

Attempts at treating subjective tinnitus with acupuncture have been disappointing and have not, as yet, demonstrated any clear benefit (87; 94), but some studies have suggested a short-term benefit of statistical significance but of unclear clinical significance (45).