GLP-1 and Tinnitus: The FDA Reports + Postmarketing Evidence

Patients on Ozempic, Wegovy, Mounjaro, and Zepbound have submitted MedWatch reports describing new-onset ringing, buzzing, or fullness in the ears, and the FDA has logged those reports in FAERS. The two published database analyses we have so far (Khan 2025 Laryngoscope^[1] and Pak 2025 Otology & Neurotology^[2]) describe what is in the database — they do not establish causation. The registrational trials (SURMOUNT-1, STEP-1, SUSTAIN, PIONEER) did not adjudicate tinnitus as a separate adverse event, and the current FDA labels for Wegovy, Ozempic, Mounjaro, and Zepbound do not list tinnitus in the warnings or precautions. Meanwhile, baseline adult tinnitus prevalence is ~14% (Jarach 2022 JAMA Neurology^[3]). For a patient who develops new ringing on a GLP-1, the right move is an ENT workup, not an internet diagnosis — and that workup catches the cheap-fix causes (cerumen, ototoxic NSAIDs, blood pressure swings) far more often than it finds a true drug effect.

The honest summary

• FAERS contains GLP-1 tinnitus reports. Khan 2025^[1] and Pak 2025^[2] both queried the FDA Adverse Event Reporting System and FDA otolaryngology adverse-event data and identified tinnitus among the reported ear-related events on semaglutide, liraglutide, dulaglutide, and tirzepatide. Both papers explicitly frame the findings as signal generation, not causation.
• The FDA labels are silent on tinnitus. The current prescribing information for Wegovy, Ozempic, Mounjaro, Zepbound, Saxenda, and Rybelsus does not list tinnitus in the warnings, precautions, or adverse-reaction sections. That silence reflects the trial datasets, not the FAERS database.
• Baseline tinnitus is common. Jarach 2022 JAMA Neurology^[3] pooled 113 studies and estimated global adult prevalence at ~14% (any tinnitus) with ~2% reporting severe tinnitus. GBD 2019^[4]reported hearing-loss prevalence affecting roughly 1.5 billion people. Any GLP-1 signal sits on top of an enormous baseline.
• New tinnitus on a GLP-1 deserves an ENT workup, not a stop order. The AAO-HNS clinical practice guideline (Tunkel 2014^[5]) defines the workup: history, otoscopy, audiogram with tympanometry, and MRI of the internal auditory canal if hearing loss is asymmetric. Most cases have a non-GLP-1 explanation.

What the two published FAERS analyses actually showed

Pak 2025 (Otology & Neurotology^[2]) queried the FDA Adverse Event Reporting System and an institutional database for otolaryngologic adverse events associated with GLP-1 receptor agonists. The leading reported events were eustachian tube dysfunction, ear pain, hearing loss, and tinnitus; the authors describe a possible mechanistic link through GLP-1 receptor expression in middle-ear and inner-ear tissues but explicitly note that database-derived signals cannot establish causation. Khan 2025 (Laryngoscope^[1]) ran a parallel FAERS analysis across semaglutide, liraglutide, dulaglutide, and tirzepatide and found a heterogeneous mix of otolaryngologic reports including taste disturbance, hearing impairment, vertigo, and tinnitus, again framing the findings as hypothesis-generating rather than confirmatory.

FAERS is a passive surveillance database. Reports are voluntary, the denominator (total exposed) is unknown, the causal attribution is the reporter's opinion, and the same patient can appear multiple times. Reporting rates rise with media attention — the “Ozempic side effect” news cycle of 2023–2025 produced predictable spikes in FAERS for symptoms that may or may not be drug-related. The Khan and Pak papers both name these limitations explicitly.

Why the trial datasets are silent

STEP-1 (Wilding 2021 NEJM), SURMOUNT-1 (Jastreboff 2022 NEJM), SUSTAIN-6 (Marso 2016 NEJM), and PIONEER 6 (Husain 2019 NEJM) used pre-specified MedDRA-coded adverse event categories. None of the published trial reports for the obesity or cardiovascular outcomes programs adjudicated tinnitus as a protocol-specified adverse event, and the routinely reported tables grouped ear-related symptoms under generic categories that did not break out tinnitus separately. That is normal for a trial powered for weight or cardiovascular endpoints — but it means the trial datasets cannot confirm or refute the FAERS signal.

The animal-model evidence runs in a different direction. Jiang 2025 (Military Medicine^[8]) reported that liraglutide pretreatment reduced blast-induced cochlear injury in a chinchilla model and identified functional GLP-1 receptor signaling in cochlear tissue. The mechanistic story therefore is at least bidirectional: GLP-1 receptors exist in the auditory system, but the animal data point toward protection rather than damage. That makes the FAERS tinnitus signal harder to explain mechanistically and easier to attribute to indirect pathways (blood pressure swings, volume depletion, sympathetic activation).

The mechanistic hypotheses (none proven)

• Vascular and pulse-related. Pulsatile tinnitus is sensitive to blood pressure, heart rate, and arterial compliance changes. GLP-1 agonists modestly raise resting heart rate and lower blood pressure during titration; a patient with subclinical pulsatile tinnitus could perceive the symptom emerging.
• Volume depletion and dehydration. Nausea, decreased fluid intake, and occasional vomiting during dose escalation can produce mild volume depletion and orthostatic symptoms — both can transiently worsen tinnitus.
• Concurrent ototoxic medications. Patients starting GLP-1s often add or increase NSAIDs (joint pain), aspirin, loop diuretics (heart failure or weight-loss edema), or aminoglycosides (acute infections) — all are documented ototoxins. The GLP-1 may be a coincident exposure, not the cause.
• Sympathetic activation. Anxiety, sleep disruption from injection-site discomfort, and stimulant appetite suppressants taken concurrently can all raise central gain in the auditory cortex, which is one of the dominant models for tinnitus generation.
• No direct cochlear receptor damage evidence. The animal data we have (Jiang 2025^[8]) point toward protection, not damage. No published study has demonstrated direct GLP-1-induced hair-cell or stria-vascularis injury.

Magnitude: putting the FAERS signal in context

The differential diagnosis the ENT will actually run

AAO-HNS guideline (Tunkel 2014^[5]) and standard otolaryngology practice produce the same short list for new unilateral or bilateral tinnitus:

• Cerumen impaction. The cheapest, most common, most missed cause. Otoscopy and a 5-minute irrigation often fix the problem.
• Noise exposure. Recent loud-music or power-tool exposure causes acute sensorineural threshold shifts with tinnitus.
• Ototoxic medications. High-dose aspirin or NSAIDs, loop diuretics (furosemide, bumetanide), aminoglycoside antibiotics, platinum chemotherapies, and quinine. The cross-check against the patient's full medication list is mandatory before blaming the GLP-1.
• Meniere's disease. Episodic vertigo, fluctuating low-frequency hearing loss, aural fullness, and tinnitus. Diagnosed by audiogram pattern and clinical course.
• Acoustic neuroma (vestibular schwannoma). Suspected for asymmetric sensorineural hearing loss on audiogram; confirmed by MRI of the internal auditory canal.
• Temporomandibular joint dysfunction. Tinnitus that worsens with jaw movement and improves with bite splints.
• Cervical spine and myofascial sources. Somatic tinnitus that modulates with neck position.
• Blood pressure swings. A new antihypertensive, an unmasked thyroid disorder, or dehydration during GLP-1 titration can all produce pulsatile tinnitus.

The workup, in order

1. Otoscopy and cerumen check. Done in the primary-care visit. If wax is impacted, irrigate and reassess.
2. Audiogram with tympanometry. The single highest-yield test. Confirms or rules out sensorineural hearing loss; tympanometry detects middle-ear effusion or eustachian tube dysfunction.
3. MRI of the internal auditory canal if the audiogram shows asymmetric sensorineural hearing loss (rule out acoustic neuroma).
4. Blood pressure trend and orthostatic check. Especially during GLP-1 titration.
5. Basic labs: thyroid stimulating hormone, vitamin B12, lipid panel, and a comprehensive metabolic panel. The yield is modest but the tests are cheap.
6. Medication reconciliation. NSAIDs, high-dose aspirin, loop diuretics, recent aminoglycoside exposure, quinine, and any new antihypertensive.

What the GLP-1 prescriber should do

• Do not stop the GLP-1 reflexively. The baseline incidence of new tinnitus is high enough that most patients who develop it on a GLP-1 would have developed it regardless.
• Hold the next dose escalation if symptoms appear during titration. Reassess in four weeks. Many symptoms attributed to escalation resolve at a stable dose.
• Push hydration. 2–3 L water per day during titration; cut concurrent NSAID and high-dose aspirin use where clinically possible.
• Refer to ENT and audiology. The audiogram is diagnostic; the ENT visit catches the cheap-fix causes.
• Keep a symptom log. Date of onset, severity (1–10), laterality, modulators (jaw position, neck position, blood pressure readings), and concurrent medications.

Treatment for confirmed chronic tinnitus

The AAO-HNS guideline (Tunkel 2014^[5]) gives the strongest recommendation to cognitive behavioural therapy for bothersome chronic tinnitus, and the Cochrane review (Fuller 2020^[6]) confirmed CBT reduces tinnitus-related distress and quality-of-life impact, although it does not consistently reduce perceived loudness. Hearing aids are first line when tinnitus accompanies a measurable hearing loss. Sound therapy (white noise, notched music, hearing-aid masking) and tinnitus retraining therapy are reasonable options. The guideline explicitly recommends against routine use of Ginkgo biloba, melatonin, zinc supplements, and transcranial magnetic stimulation for primary tinnitus management. The Cochrane-level review of Ginkgo biloba (Kramer 2018 Medwave^[7]) found no clinically meaningful benefit.

Antidepressants (sertraline, nortriptyline) have weak evidence and are reserved for comorbid depression or anxiety. Benzodiazepines should be avoided for chronic tinnitus because the tolerance and dependence risk outweighs the short-term benefit.

Insurance and cost

ENT evaluation is covered under standard commercial and Medicare benefits when ordered for a documented otologic symptom. Audiogram with tympanometry runs roughly $100–300 out-of-pocket if uninsured. Hearing aids are not covered by standard Medicare Part B (Medicare Advantage plans often include a hearing benefit; many state Medicaid programs cover them); out-of-pocket cost for a pair is typically $1,500–6,000. Over-the-counter hearing aids (FDA category since 2022) start around $300–1,500 for adults with mild-to-moderate loss.

Related research and tools

• GLP-1 side effects: fatigue and hair loss timeline — the broader pattern of non-GI symptoms on Wegovy and Zepbound
• GLP-1 and brain fog — the cognitive-symptom side of central nervous system effects, what FAERS shows, and the workup
• GLP-1 heart rate and palpitations — the cardiovascular signal that overlaps with pulsatile tinnitus reports
• GLP-1 and diuretics — loop diuretics are independently ototoxic; the interaction with GLP-1-driven dehydration matters
• GLP-1 side effect questions answered — the patient-facing FAQ hub
• GLP-1 side effect timeline tool — map symptom onset against dose-escalation week

Important disclaimer. This article is educational and does not constitute medical advice. New-onset tinnitus, particularly if asymmetric, accompanied by hearing loss, vertigo, or neurologic symptoms, deserves prompt otolaryngology evaluation regardless of GLP-1 status. FAERS-based signal analyses describe what was reported, not what was caused. The FDA labels for Wegovy, Ozempic, Mounjaro, Zepbound, Saxenda, and Rybelsus did not list tinnitus in the warnings or precautions sections as of the last verification date. PMIDs were verified live against the PubMed E-utilities API on 2026-05-29.

Last verified: 2026-05-29. Next review: every 12 months, or sooner if a prospective trial with adjudicated otologic endpoints, an updated FDA label, or a Bradford-Hill style analysis of the FAERS signal is published.