The tinnitus studies most worth knowing about since 2022 are, newest first: a 2026 field-wide overview, the 2025 UNITI trial of combined treatments, a 2025 estimate of the change patients notice, Lenire's pivotal trial and the largest genetic study yet.
Each entry says what kind of study it was, what it found, how much weight it can bear, and which article covers it in full. Entries are dated by when the paper was first published, online where that came first. When a later result changes how an earlier entry reads, the entry says so. For the general method behind the "weight" lines, see the patterns that make a weak tinnitus result look strong.
2026
May 2026 — The Nature Reviews Disease Primer on tinnitus
Ten researchers wrote the 2026 Nature Reviews Disease Primers volume on tinnitus, a review of the whole field rather than a new study. It puts tinnitus at about 14% of adults, with about 2% severely affected. It names tinnitus-focused counselling and cognitive behavioural therapy as first-line treatments, with hearing rehabilitation and treatment of somatosensory contributors as valuable adjuncts, and says emerging neuromodulation, bimodal stimulation included, benefits selected subgroups. Its list of new research areas: biomarkers, tinnitus genetics, inner ear regeneration, closed-loop neuromodulation and digital therapeutics.
Weight: an expert review, not new data. Its value is as a statement of where researchers across the field place the evidence.
2025
November 2025 — UNITI: do combined treatments beat single ones?
UNITI randomised 461 people with chronic tinnitus, at five clinical sites in the EU, to one treatment or a combination of two for 12 weeks. The treatments were CBT, hearing aids, app-based structured counselling and app-based sound therapy, making ten arms in all; the app-based ones were delivered without contact with a clinician.
On the Tinnitus Handicap Inventory, scores fell by 14.9 points with combination treatment against 11.7 with a single one (p = 0.034). The trial found no synergy. CBT and hearing aids alone had large effect sizes that combining could not increase, and the authors describe a compensatory effect, in which a more effective treatment offsets a less effective one. No serious adverse events occurred.
Weight: a large randomised trial. The advantage for combinations is 3.2 points between the arms, and the authors attribute it to the stronger treatment in each pair rather than to two treatments adding together. Combining tinnitus treatments covers what that means for choosing.
November 2025 — Deep brain stimulation, tested for safety in four people
A feasibility study from Maastricht University Medical Centre, published online in Neurotherapeutics, implanted electrodes in the medial geniculate body, a relay station on the auditory pathway, in four people with severe, chronic, treatment-refractory tinnitus. The design was a randomised, double-blind crossover: stimulation on then off, or off then on. There were no irreversible stimulation-induced side effects, surgical side effects resolved within two weeks, and three of the four improved on the Tinnitus Functional Index.
Weight: four people, in a study built to test safety rather than effectiveness, which its authors say needs a follow-up study. Where tinnitus research is headed sets out what it does and does not show.
October 2025 — Gene therapy restores hearing in children with OTOF deafness
A trial of the gene therapy DB-OTO, published online in the New England Journal of Medicine, treated twelve children with OTOF-related deafness, an inherited condition in which the hair cells are present but cannot release their signal. Nine of the twelve met the primary hearing endpoint at 24 weeks, six could hear soft speech without assistive devices, and three reached average normal hearing sensitivity. It followed a 2024 Lancet trial of a similar therapy in which five of six children recovered hearing.
Weight: a real result for a rare genetic condition. Neither trial measured tinnitus, and neither regenerated hair cells: the therapy switches existing cells back on, which is not possible where noise or age has destroyed them. Hair cell regeneration and gene therapy explains the difference.
May 2025 — What predicts tinnitus, in 192,993 people
A machine-learning analysis of UK Biobank data in Nature Communications used socio-demographic, psychological and health information from 192,993 participants, 41,042 of them with tinnitus, to predict who has tinnitus and how severe it is. Hearing health was the main risk factor for both. Mood, neuroticism and sleep predicted severity, and the severity model predicted how tinnitus progressed over nine years.
Weight: very large and observational. It identifies predictors, not causes. The UK Biobank tinnitus study covers the prevalence figures from the same analysis.
March 2025 — How big a change in THI or TFI score patients notice
A multi-centre study in Otolaryngology–Head and Neck Surgery, run across European tinnitus centres, estimated the smallest change patients notice by anchoring score changes to patients' own ratings of being minimally improved. After 12 weeks it came out at 9 points on the Tinnitus Functional Index (N = 359) and 11 points on the Tinnitus Handicap Inventory (N = 364). Both thresholds were larger for people who started more severe, and over longer intervals.
Weight: the most direct estimate available of what a score change means to the person reporting it, and a reminder that there is no single number. How tinnitus is measured shows how to use it.
2024
August 2024 — TENT-A3, the trial behind Lenire's authorisation
Lenire pairs sound through headphones with electrical stimulation of the tongue. Its pivotal trial, TENT-A3, was single-arm: all 112 participants had six weeks of sound alone and then six weeks of bimodal treatment, always in that order, each serving as their own control. A responder was someone whose Tinnitus Handicap Inventory score fell by more than 7 points.
In the 44 participants still moderately bothered after the sound-only weeks, 58.6% responded during the bimodal weeks against 43.2% during the sound-only weeks. In the full cohort, that advantage was not observed. The authors write that a sham control was not possible, because both the sound and the tongue stimulation are noticeable. The device was granted marketing authorisation through the FDA De Novo pathway in March 2023.
Weight: the fixed order cannot separate the tongue stimulation from time, expectation or order effects, and no sham-controlled trial of the device exists.
What followed. After the 2023 authorisation, a retrospective chart review from one US clinic, published in April 2025, reported a 91.5% responder rate among patients with bothersome tinnitus after about 12 weeks. It is an uncontrolled, retrospective chart review: no comparison group, no sound-only stage, and every patient in it received tinnitus counselling and education before choosing a treatment plan. The pivotal trial's own 12-week figure, counted from the start of treatment, was 79.4% of its full cohort — also a before-and-after number with no control. A chart review with no sound-only stage has nothing to set the tongue stimulation against, so it cannot confirm that advantage. The chart review's first author declares being a commercial provider of the treatment, and a co-author declares financial interests in the manufacturer. Bimodal neuromodulation devices covers all three Lenire trials and who funded them.
January 2024 — The largest genetic study of tinnitus
A genome-wide association study in Nature Communications combined data from UK Biobank and the US Million Veteran Program, 596,905 people in its cross-ancestry analysis, and identified 39 tinnitus risk loci. It confirmed many genetic variants shared with hearing difficulty, but also found a distinct genetic architecture for tinnitus, with higher polygenicity and a large proportion of variants not shared with hearing difficulty. Tinnitus was also genetically correlated with a range of psychiatric disorders, major depressive disorder among them.
Weight: strong evidence that tinnitus is shaped by many genes of small effect and is genetically more than a by-product of hearing loss. It produced no test and no treatment. Genetics and tinnitus covers the twin and family studies behind it.
2023
June 2023 — OTO-313 does not beat a placebo injection
A randomised, double-blind, placebo-controlled phase 2 trial of OTO-313, a slow-release form of the NMDA antagonist gacyclidine injected through the eardrum, found no significant benefit over placebo in people with moderate to severe one-sided tinnitus of two to twelve months' duration. The authors attribute that to an unexpectedly high placebo response, particularly among patients with chronic tinnitus, despite training designed to reduce it.
Weight: a clean negative result, and a caution for every uncontrolled tinnitus study: a placebo injection, in a trial designed to limit exactly this, did as well as the drug. The tinnitus drugs you read about follows it and the other candidates that have reported.
2022
September 2022 — How many people have tinnitus
A systematic review and meta-analysis in JAMA Neurology pooled 83 prevalence studies and put tinnitus at 14.4% of adults: more than 740 million adults worldwide, of whom more than 120 million experience it as a major problem. Severe tinnitus came out at 2.3%.
Weight: the anchor figure for tinnitus statistics. National estimates differ from it mainly because of how the question is asked, which global tinnitus prevalence explains.
What this log leaves out
Announcements, press releases, and trials that have not yet reported. A registered trial is a question rather than a result; how to find and join a tinnitus clinical trial covers where to look for the ones recruiting. An earlier version of this page described the Lenire chart review above as confirmation of the trial, and listed the UNITI trial as ongoing after it had reported. The first is recorded on the corrections page; both entries now say what the studies show.