The Tinnitus Drugs You Read About: What Happened to Them

Search almost any tinnitus drug and you will find an article announcing a promising new treatment. Search the same compound's trial result and you will usually find nothing, because a failed trial gets a journal paper and a successful press release gets everything else.

This page is the other half. Three compound names people actually search for, and what each one did when it was tested.

The standing fact

There is no approved drug for tinnitus. Not a weak one, not a partially effective one — none.

The most useful place to read that is in a 2022 review in Hearing Research written by the developers of one of the candidates on this page, who had every commercial reason to phrase it optimistically and did not: no therapeutics for any auditory related indication have been FDA approved.

What drugs do have a role covers the repurposed medications — anticonvulsants, lidocaine, and where a prescription genuinely helps. This page is about the ones developed specifically for this.

AM-101, and the shape of a hopeful result

AM-101 is an NMDA receptor antagonist delivered by injection through the eardrum, aimed at tinnitus arising soon after acoustic trauma, sudden hearing loss or middle ear infection — the idea being to interrupt glutamate excitotoxicity in the cochlea before the tinnitus becomes established.

Its phase II trial, published in Otology & Neurotology in 2014, was serious work: 248 patients, 28 European sites, double-blind, placebo-controlled, three injections over three consecutive days.

The study overall failed to demonstrate a treatment benefit on its primary endpoint, which was change in minimum masking level.

Then the rest of the result, which is why the compound stayed famous. The higher dose showed statistically significantly better improvement for tinnitus loudness, annoyance, sleep difficulties and tinnitus impact — in the subgroup whose tinnitus followed noise trauma or otitis media. The sudden-hearing-loss subgroup was inconclusive. The authors concluded the study established proof of concept, and argued that patient-reported outcomes are more relevant and reliable for tinnitus than psychoacoustic tests.

That last argument may well be right. It is also the argument you make when your psychoacoustic primary endpoint has failed and your patient-reported secondaries have not.

This is the single most useful pattern to be able to recognise in this field: a failed primary endpoint, a positive subgroup, and a conclusion of proof of concept. It is not fraud and it is not meaningless — subgroups do sometimes identify the population a drug works in. It is also exactly what a promising press cycle is built from, and a subgroup finding is a hypothesis for the next trial rather than a result. How this site reads evidence sets out the general version.

A 2025 review of NMDA receptors as tinnitus targets lists AM-101 alongside acamprosate, caroverine and neramexane as antagonists whose trials have shown promising results — and adds, in the same sentence, that none are yet approved. Four compounds, one mechanism, decades of work, no prescription.

OTO-313, and a placebo response nobody could design around

OTO-313 is a sustained-release formulation of gacyclidine, also an NMDA antagonist, also given as a single injection through the eardrum. Its phase 2 study was published in the European Archives of Oto-Rhino-Laryngology in 2023: randomised, double-blind, placebo-controlled, in patients with moderate to severe one-sided tinnitus of two to twelve months' duration, followed for sixteen weeks.

It did not beat placebo. Similar proportions of responders on the Tinnitus Functional Index at weeks 4, 8, 12 and 16. Similar daily ratings of loudness and annoyance. Similar global impression of change. No significant difference in any pre-specified stratum, though the numbers favoured OTO-313 in the group whose tinnitus was under six months old.

The explanation the authors give is the part worth carrying away. They report an unexpectedly high placebo response, particularly among patients with chronic tinnitus — despite training implemented specifically to mitigate it.

Think about what that means for everything else you read. A group of people receiving an injection of nothing, in a trial designed by people who anticipated this exact problem and tried to prevent it, still improved enough to erase the drug's effect. Tinnitus fluctuates, attention to it fluctuates, and being enrolled in a trial changes both. Any uncontrolled study of any tinnitus treatment is measuring that alongside whatever it thinks it is measuring — which is why an uncontrolled positive result is worth so much less than it sounds.

The drug was safe and well tolerated. It simply did nothing a placebo injection did not.

SPI-1005, which is not quite what it is described as

SPI-1005 is ebselen, an anti-inflammatory and neuroprotective compound, and unlike the two above it is taken orally and is still in active development. It comes up constantly in tinnitus discussion.

What the published record actually says, from a 2022 review in Hearing Research: ebselen reduced noise-induced and aminoglycoside-induced hearing loss in animals; multiple phase 2 trials have demonstrated safety and efficacy in Ménière's disease and acute noise-induced hearing loss; and it was being tested to prevent and treat tobramycin-induced ototoxicity in cystic fibrosis patients.

Read that list for what is not in it. The programme is about protecting and preserving hearing. Tinnitus appears in the review's title and in the conditions the company studies, but the phase 2 efficacy claims are for Ménière's disease and acute noise damage — not for chronic subjective tinnitus, which is what most people reading about it have.

One thing to weigh when reading it: both authors of that review work for the company developing the drug. That does not make it wrong, and it is a legitimate publication. It does mean the sentence about no approved therapeutics carries more weight than the sentences about promise, because the first runs against the authors' interest and the second runs with it.

What the record adds up to

• Every compound developed specifically for tinnitus that has reported a controlled phase 2 result has failed its primary endpoint.
• The failures have been informative. The OTO-313 placebo finding constrains how every future tinnitus drug trial has to be designed.
• The mechanism most invested in — NMDA receptor antagonism — has four named compounds and no approval.
• The one programme still visibly active is aimed at protecting hearing rather than treating established tinnitus.

None of that means a drug will never work. Where tinnitus research is headed covers what is being tried now, and can tinnitus be cured covers why the ordinary case is so resistant while a short list of structural causes can be fixed outright.

What to do with this

• If you found a hopeful article about a tinnitus drug, check the year, then search the compound name with "phase 2" or "randomized". The trial result is usually findable and usually less exciting than the announcement.
• If a result is described as proof of concept, look for the primary endpoint. AM-101's is the textbook case: the phrase is doing work that the primary result did not.
• If a study had no placebo group, the OTO-313 finding is the reason to discount it heavily. Placebo response in chronic tinnitus was strong enough to defeat a real drug in a trial built to prevent exactly that.
• Do not wait for a drug. Nothing on this page is close, and the treatments with evidence are available now.
• Trials still need participants, and joining one is a reasonable thing to do with clear eyes — how to join a tinnitus clinical trial covers what that involves.