Tinnitus: Independent Evidence on Causes, CBT, Devices & Cures

Table of contents

• Evidence summary
• What tinnitus is
• Symptoms and red flags
• Causes and risk factors
• Treatments that work, and what does not
• Medicines
• Supplements and lifestyle claims, graded
• Risks and harms
• Who to see and what to expect
• Follow the money
• Related research
• Frequently asked questions
• Sources and funding notes

Evidence summary

Independent evidence and credibility scorecard

Independence tiers: 1 = regulator or government body; 2 = academic, guideline or Cochrane team with no declared commercial funding for the work; 3 = independent authors analysing mostly manufacturer-run trials, or minor declared industry ties; 4 = manufacturer-funded or manufacturer-authored. Credibility runs from A (highest) to D. For tinnitus, a charity that takes sponsorship from device makers is placed in tier 3. Our sources cluster in the UK and US, with Dutch, Brazilian and Irish contributions, so the verdict reflects mainly Anglo-American practice.

What tinnitus is

Tinnitus is hearing a sound that has no outside source. The NHS describes ringing, buzzing, humming, hissing or even music, in one ear, both ears or the head (NHS). The Cochrane CBT review uses a working definition of chronic tinnitus as lasting at least three months (Fuller et al.). The 2014 US guideline separates tinnitus that bothers a person and lasts six months or more from tinnitus that does not bother them, because management differs (AAO-HNSF).

It is not one disease. NICE uses terms that matter in the clinic. Subjective tinnitus is heard only by the person. Objective tinnitus is noise generated in or near the ear that an examiner can detect, and it is less common (NICE NG155, terms used). Pulsatile tinnitus beats with the heartbeat and has a different set of possible causes from steady ringing.

How common is it?

The best pooled figure is 14.4% of adults (95% CI 12.6% to 16.5%), with wide variation between studies (4.1% to 37.2%). It rises with age: 9.7% at 18 to 44, 13.7% at 45 to 64 and 23.6% at 65 or over. Severe tinnitus is far rarer, at 2.3% (Jarach et al. 2022). That review was funded by European Union Horizon 2020 grants and an Italian tinnitus association, and the funders had no role in the analysis, but one author reported personal fees from a tinnitus device company and a supplement maker (disclosures).

Symptoms and red flags

The sound itself varies in pitch and loudness. The NHS lists effects on sleep, concentration, mood and anxiety, and advises seeing a GP if tinnitus is constant, getting worse or bothering you (NHS). NICE tells clinicians to ask about sleep and to screen for insomnia (for example with the Insomnia Severity Index) and to stay alert to low mood and anxiety (NG155, 1.3.5 to 1.3.8).

Equally important, imaging is not advised for everyone. Both guidelines advise against scans for steady, two-sided, non-pulsatile tinnitus with no other signs (NICE 1.4.8; AAO-HNSF). If you read a secondary summary that places persistent pulsatile or one-sided tinnitus in an urgent bracket, check it against the NICE text above, which uses "consider referring" for those two features when there are no other signs.

Causes and risk factors

The cause is often unclear. The NHS links tinnitus to hearing loss, Ménière's disease, conditions such as diabetes, thyroid disorders and multiple sclerosis, anxiety or depression, and some medicines including certain chemotherapy drugs, antibiotics, NSAIDs and aspirin (NHS). Earwax and ear infections are the treatable causes a GP checks first. NICE says tinnitus is commonly associated with hearing loss and not commonly with other physical problems, that stress and loud noise can make it worse, and recommends safe-listening advice (NG155, 1.1.3 and 1.1.4). NICE lists an audiological assessment as the core investigation for everyone with tinnitus (1.4.1).

If you take aspirin or other regular medicines, do not stop them on your own because of tinnitus; ask a prescriber. For background on aspirin and blood thinners see our guides on aspirin and anticoagulants and bleeding safety.

Treatments that work, and what does not

CBT in detail

The Cochrane review pooled 28 trials (2,733 adults, mostly European, average age 43 to 70, CBT lasting 3 to 22 weeks). Against waiting list, CBT may lower tinnitus impact (SMD -0.56, 95% CI -0.83 to -0.30; low certainty), equal to 10.91 points on the Tinnitus Handicap Inventory, against a smallest noticeable change of about 7. Against ordinary audiological care the gap was smaller, 5.65 points (95% CI -9.79 to -1.50; 3 studies, 444 people; moderate certainty). Against other active treatments such as relaxation it was 5.84 points. CBT may slightly lower depression (SMD -0.34, 95% CI -0.60 to -0.08), while effects on anxiety are uncertain. No serious adverse effects were reported in the studies that tracked them (Fuller et al. 2020).

The largest single trial, in the Netherlands, gave 245 people stepped CBT-based specialist care and 247 usual care. Over 12 months the specialist group did better on quality of life (difference 0.059, 95% CI 0.025 to 0.094), tinnitus severity (d = 0.43) and impairment (d = 0.45), with no adverse events. That trial was funded by the Netherlands Organisation for Health Research and Development (Cima et al., Lancet 2012). Its package also included sound-focused retraining, so the CBT part cannot be separated out.

For access, a UK trial randomised 92 people to eight weeks of guided internet CBT or face-to-face NHS care. Tinnitus Functional Index scores differed by 5.18 points (95% CI -4.17 to 14.53) at the end, within the 13-point margin chosen to call the internet version "not worse"; only insomnia outcomes fell outside the margin, in favour of the internet course (Beukes et al. 2018). The sample was small, which is why the confidence interval is wide. See our wider guides on CBT and digital CBT.

Hearing aids and sound therapy

The 2014 Cochrane review of hearing aids for tinnitus with hearing loss found only one trial (91 people) comparing hearing aids with sound generators, and the review authors declared research with a hearing-aid maker and a tinnitus clinic (Hoare et al. 2014). The 2018 review of sound therapy found 8 studies (590 people), none comparing devices with waiting list, placebo or information alone, and concluded there was no evidence of superiority over those comparators. Combination hearing aids did not beat plain hearing aids (SMD -0.15, 95% CI -0.52 to 0.22; 3 studies, 114 people). Both device types were linked to a clinically meaningful drop in symptom score over time, which is hard to interpret without a no-device control (Sereda et al. 2018).

Tinnitus retraining therapy

The US Tinnitus Retraining Therapy Trial randomised 151 military personnel, retirees and dependents to full retraining (counselling plus sound generators), counselling with placebo sound generators, or standard care. At 18 months there was no difference in Tinnitus Questionnaire scores between groups. All improved: the effect size was -1.32 (95% CI -1.78 to -0.85) for full retraining, -1.16 for the partial version and -1.01 for standard care (Scherer et al. 2019). The US NIH institute that funds deafness research paid for it. The shared improvement may reflect time, attention or placebo effects, which the trial design cannot separate from true benefit.

Bimodal devices: Lenire

Lenire pairs headphone sound with mild electrical pulses on the tongue. The US FDA record shows a De Novo request from Neuromod Devices Limited of Dublin, received 18 August 2021 and granted on 6 March 2023, under product code QVN (FDA database). A De Novo grant authorises a new type of lower-risk device for sale. It is not proof of a cure, and it is not the same as a trial showing superiority over a no-device control.

The pivotal trial (TENT-A3) enrolled 112 people, each acting as their own control: six weeks of sound alone followed by six weeks of bimodal treatment. In people with moderate or worse tinnitus, 58.6% (95% CI 43.5% to 73.6%) were responders on the Tinnitus Handicap Inventory after bimodal treatment against 43.2% (95% CI 29.7% to 57.8%) after sound alone (p = 0.022); on the Tinnitus Functional Index it was 45.5% against 29.6% (p = 0.010). The edge was not seen in the full cohort across all severity levels. There were no device-related serious adverse events. The authors state that a sham control was not possible because both stimuli are noticeable, and that the study was funded by the sponsor, Neuromod Devices; one author is a paid consultant to the firm (Boedts et al. 2024; full text). An earlier randomised, double-blind trial's primary comparison was between two bimodal settings, not bimodal against none (it also had two exploratory arms, one of which gave sound alone first), and its authors include company employees, shareholders and paid advisers (Conlon et al. 2022; competing interests).

A patient charity that lists the maker as a corporate member and past sponsor says treatment is 60 minutes daily for at least 12 weeks, costs around £3,600, and that in the trials about a third of people reported temporary worsening of tinnitus. It also says independent trials with a waiting-list control or similar are needed (Tinnitus UK, 2024). We found no independent trial and no NICE recommendation. We could not verify its UK regulatory status.

Medicines

NICE makes one drug recommendation for tinnitus itself: do not offer betahistine (NG155, 1.5.5). The US guideline advises against antidepressants, anticonvulsants, anxiolytics or intratympanic (ear-injected) drugs for routine treatment of persistent, bothersome tinnitus, and against transcranial magnetic stimulation (AAO-HNSF 2014).

The 2012 Cochrane review of antidepressants found six trials (610 patients) of generally low quality. Four tricyclic trials suggested slight improvement that might reflect bias, and the one SSRI trial found no overall improvement (Baldo et al. 2012). A later network meta-analysis of 36 trials (2,761 people) ranked oral amitriptyline highest for severity, but quality of life did not differ from placebo and the authors only suggested a "potential role" (Chen et al. 2021). Treating depression, anxiety or insomnia that goes with tinnitus is reasonable; for those uses see our guides on major depression, generalised anxiety and amitriptyline.

Supplements and lifestyle claims, graded

Risks and harms

• Missed causes. Sudden hearing loss is time-critical, and NICE sets a 24-hour window (1.2.3). Delay in an urgent case is the main avoidable harm.
• Mental health. The NHS names anxiety and depression as both a cause and a consequence (NHS). Persistent distress, panic, hopelessness or thoughts of self-harm should be raised with a clinician (see the crisis note above).
• Treatment harms are mostly low. CBT trials reported no serious adverse effects in those that tracked them (Cochrane). The Lenire pivotal trial reported no device-related serious adverse events, while the charity summary reports temporary worsening in about a third of trial participants (Boedts; Tinnitus UK).
• Financial harm. Heavily marketed formulas are sold at a premium over equivalent vitamin mixtures, and a market review found all of those examined made unfounded claims of relief (Vendra et al. 2019).

Who to see and what to expect

Start with a GP, who looks for earwax or infection, checks hearing, and may refer to audiology or ENT (NHS). NICE advises an audiological assessment for everyone with tinnitus, tympanometry when middle-ear problems are suspected, and a short questionnaire such as the Tinnitus Functional Index to gauge impact. It advises against routine acoustic reflex testing, loudness discomfort testing and pitch-and-loudness matching (NG155, 1.3 and 1.4). If distress is high, a psychologist or an audiology service offering CBT is the next stop. Children and young people fall under the same guideline, with age-appropriate measures. For overlapping sound sensitivity see our guide to misophonia; for tinnitus with hearing loss in an inflammatory disease see Cogan syndrome.

Follow the money: who makes it and who funded the evidence

Who sells tinnitus products

The market has three layers. First, hearing-aid and sound-generator clinics sell devices and sound programmes, many privately. We did not trace the ownership of hearing-aid makers or private tinnitus clinics, and could not find a citable market-size figure, so we state none. Second, over-the-counter formulas. A peer-reviewed review found a "robust and diverse market" of unproven products, mostly mixtures of inexpensive common vitamins, minerals or herbs sold at a premium over similar preparations not advertised for tinnitus. It named two brands that advertise in ENT journals and feature supposed ear, nose and throat doctor endorsements: Arches Tinnitus Formula (Arches Natural Products, Salt Lake City) and Lipo-Flavonoid (Clarion Brands, Solon, Ohio) (Vendra et al. 2019). Third, prescription-only or clinic-fitted neuromodulation, led by Lenire.

Lenire's maker and backers

Neuromod Devices Limited is based in Dublin, Ireland, and is the requester on the FDA record (FDA). Its own release says it raised €10.5 million in a Series B in October 2020 led by Fountain Healthcare Partners, described as the majority investor, with existing investors Moffett Investment Holdings and Medical Device Resources and venture debt from Kreos Capital and Silicon Valley Bank (BioSpace/Business Wire). A 2023 announcement describes a further €30 million: a €15 million Series B expansion led by Panakès Partners and €15 million of venture debt from the European Investment Bank, aimed at launching in the US (Hearing Health Matters). These are company statements, not audited accounts, and we did not trace individual stakes. We also saw a reference to a 2025 raise but did not open it.

Who funded the trials

Public and charity money dominates the CBT and retraining evidence: NIHR infrastructure for the Cochrane reviews, a Dutch public health research fund for the Lancet CBT trial, the US NIH deafness institute for the retraining trial, and a UK Department of Health grant in PubMed's record for the internet CBT trial. By contrast the device evidence is sponsor-funded: the pivotal Lenire trial was funded by Neuromod Devices, with its consultant among the authors, and the randomised settings trial lists company employees and shareholders (Boedts; Conlon). PubMed lists a Wellcome Trust grant against both papers, which does not match the funding statement in the full text (the full texts name the Wellcome clinical research facility in Dublin as a study site, and we could not resolve the discrepancy). Several tinnitus researchers appear in multiple roles. One Cochrane CBT author declared consultancy for firms including Neuromod, another reported research with a hearing-aid maker, and a prevalence review author reported personal fees from Neuromod and Schwabe, and three authors of the 2022 Lenire settings trial (including one who also co-authored the 2014 and 2018 Cochrane device reviews, which pre-date that declaration and list no Neuromod tie) serve on Neuromod's paid clinical advisory board (Fuller; Hoare; Jarach). Conflicts are disclosed, which is what we want, and they are reasons to want independent replication, not evidence of error.

Documented integrity events and "cure" marketing

• FTC, 2020. In a stipulated order, a group of related US and Canadian companies and two principals agreed to a $38,183,530 judgment, suspended after payment of $1,300,000, and to a ban on claiming that covered products eliminate or significantly reduce pain, tinnitus and other conditions without human clinical evidence. The covered products named in the order were Neurocet, Regenify and Resetigen-D (court order; FTC press release). The press release itself does not mention tinnitus, and the order was agreed by stipulation rather than after a trial, so this is a documented but indirect example.
• FTC, 2005. A complaint filed in a US federal court alleged that the supplements Arcomig and Essiac Formula were marketed as treating or curing serious conditions including cancer, diabetes, lupus and tinnitus. This is an allegation; we did not verify the outcome (FTC complaint).
• Advertising self-regulation, 2016. After a National Advertising Division inquiry, a National Advertising Review Board panel found support for saying Lipo-Flavonoid Plus "may" give relief to some people with tinnitus, but not for stronger claims such as significant or complete relief, and recommended dropping the "number one ear doctor recommended" relief claim (it may still say it is the top doctor recommendation as an adjunct therapy). The maker, Clarion Brands, said it would accept the decision (BBB National Programs).
• What we did not find. We searched for a tinnitus-specific FTC press release and an FDA warning letter against a tinnitus supplement and found neither. That is a limit of our search, not proof that none exists.

How to read the funding picture

The funding does not split neatly into good and bad. The best advice on CBT and sound therapy comes from public funders whose reputations depend on accuracy, and a placebo-controlled public trial was able to deflate a popular therapy. The devices and supplements with the most marketing have the weakest or most conflicted evidence. Even so, public funders have their own pull toward positive findings, because tinnitus researchers depend on grants and charities in a small field.

Related research

For wider, evidence-graded context, see these Pure City Research guides:

• Cognitive behavioural therapy (CBT)
• Digital CBT and mental health apps
• Mindfulness-based therapy
• Acceptance and commitment therapy
• Melatonin
• Insomnia
• CBT-I for insomnia
• Misophonia
• Generalised anxiety disorder
• Amitriptyline
• Cogan syndrome
• Arteriovenous malformations

Frequently asked questions

Sources and funding notes

What we could not verify. NICE pages blocked our automated reader, so we read the NICE recommendations through a plain web download; the NICE evidence reviews themselves were not opened. We could not open the full text of the 2014 US guideline, so its funding statement and conflicts are unchecked. We did not verify the funding models of BBB National Programs, the FDA or the NHS website. We did not confirm the outcome of the 2005 FTC complaint, find a tinnitus-specific FTC press release or FDA warning letter, trace the owners of hearing-aid makers or the makers of ginkgo extracts, or open a source for the Lenire 2025 financing. We did not confirm Lenire's UK regulatory status. We found no independent randomised trial of Lenire against a no-device control, and no source for melatonin trial funding. We did not review the individual tinnitus trials of mindfulness or ACT, or the evidence on rTMS beyond one guideline statement. The 2014 guideline predates later work.

1. NICE NG155, Tinnitus: assessment and management (2020) — guideline; sponsored by the UK Department of Health and Social Care.
2. GOV.UK page on NICE — NICE status and sponsor; UK government.
3. NHS: Tinnitus — patient information; UK public health service.
4. Fuller et al., Cognitive behavioural therapy for tinnitus, Cochrane 2020 — systematic review; NIHR infrastructure for Cochrane ENT and the Dutch ZonMW fund (full text).
5. Sereda et al., Ginkgo biloba for tinnitus, Cochrane 2022 — systematic review; NIHR and a Cochrane Complementary Medicine Field bursary (full text).
6. Person et al., Zinc supplementation for tinnitus, Cochrane 2016 — systematic review; NIHR infrastructure (full text).
7. Sereda et al., Sound therapy for tinnitus, Cochrane 2018 — systematic review; NIHR and a charity fellowship declared (full text).
8. Hoare et al., Hearing aids for tinnitus with hearing loss, Cochrane 2014 — systematic review; MRC funding listed in PubMed; authors declared research with a hearing-aid maker and a tinnitus clinic.
9. Baldo et al., Antidepressants for tinnitus, Cochrane 2012 — systematic review; no sources of support declared (full text).
10. Scherer et al., Tinnitus Retraining Therapy Trial, JAMA Otolaryngology 2019 — randomised trial; US NIDCD grants.
11. Cima et al., Lancet 2012 — randomised trial; Netherlands Organisation for Health Research and Development.
12. Beukes et al., JAMA Otolaryngology 2018 — randomised trial; UK Department of Health grant listed in PubMed; no conflicts reported.
13. Tunkel et al., Clinical practice guideline: tinnitus, 2014 — US professional society guideline; funding not seen.
14. Jarach et al., Global prevalence and incidence of tinnitus, JAMA Neurology 2022 — prevalence meta-analysis; EU Horizon 2020 grants and an Italian tinnitus association (full text).
15. Rosenberg et al., Effect of melatonin on tinnitus, Laryngoscope 1998 — small crossover trial; no funding statement seen.
16. Miroddi et al., Melatonin for tinnitus: a review, 2015 — review; no funding statement seen.
17. Menon et al., Over-the-counter treatments for tinnitus, 2026 — systematic review; authors report no funding or conflicts (full text).
18. Chen et al., Pharmacologic treatment of tinnitus, network meta-analysis 2021 — meta-analysis; reports no funding or conflicts.
19. NCCIH: Ginkgo — patient information; US National Institutes of Health.
20. Vendra et al., Over-the-counter tinnitus "cures", Laryngoscope 2019 — market review; no funding statement seen.
21. FDA De Novo database, DEN210033 — regulatory record; US government.
22. Boedts et al., Nature Communications 2024 — Lenire pivotal trial; funded by Neuromod Devices (full text).
23. Conlon et al., Scientific Reports 2022 — Lenire settings trial; authors include Neuromod employees and advisers (full text).
24. Tinnitus UK, Lenire information sheet, 2024 — patient charity; maker is a corporate member and past sponsor.
25. Neuromod Series B announcement, 2020 — company release; maker's own statement.
26. Hearing Health Matters, Neuromod €30 million financing, 2023 — trade site reproducing a company announcement.
27. FTC v. Mile High Madison Group et al., stipulated order, 2020 — court order; US government.
28. FTC press release, 20 April 2020 — regulator announcement; US government.
29. FTC v. Latinos Group Promotions et al., complaint, 2005 — complaint (allegations only); US government.
30. NARB decision on Lipo-Flavonoid Plus, 2016 — advertising self-regulation; funding model not checked.
31. Cochrane conflict of interest policy for groups — policy; Cochrane entities may not accept funds from commercial sources with a financial interest in the topic.

Educational research reviewed 6 October 2026. This article is not medical advice; talk to a qualified clinician or pharmacist before starting, stopping or combining treatments.