If you get to a doctor inside two weeks, a cheap course of generic steroids gives you the best shot at full recovery; past four to six weeks, the ear you have is the ear you keep. The action is once-only and small: go to the emergency department or a same-day ENT. The stakes are large: a permanent dead ear comes with chronic tinnitus, harder hearing in noise, and the slow downstream cost of depression and cognitive load that single-sided deafness carries.
The inner ear is unusually fragile. The hair cells that turn vibrations into nerve signals don't grow back, and they sit downstream of a single small artery with no backup supply. When anything goes wrong — a virus reactivating in the hearing nerve, a clot in that artery, the immune system attacking the inner ear by mistake — the damage starts on a clock. Days of inflammation that nobody treats become weeks of dead hair cells nobody can bring back.
In more than nine out of ten cases, no one ever finds the cause, which is why doctors call it "idiopathic." That sounds like an excuse but it shapes the treatment logic: rather than wait to know which of four mechanisms is at work, the guideline tells doctors to throw a high-dose steroid course at the inflammation, because that single drug covers most of the plausible causes well enough Chandrasekhar et al. 2019.
What the trial actually showed
About 27 in every 100,000 adults in the United States get this every year, climbing to 77 in people over 65 — far more common than the older textbooks suggested Alexander & Harris 2013. Without treatment, about two-thirds of people recover most of their hearing on their own, usually inside two weeks Mattox & Simmons 1977. That sounds reassuring until you reverse it: one in three are left with a permanently dead ear if no one acts, and the people who recover spontaneously do so almost always in the same window in which steroids work.
How much hearing comes back depends a lot on how much was lost. Mild and moderate drops, especially ones that mostly affect low pitches, recover well most of the time. Severe and profound losses — particularly when the high frequencies go — recover badly, often not at all. If the room spins at the same time the hearing goes, that's a worse signal too.
Why most people miss it
The symptom announces itself badly. It doesn't feel like "I am going deaf in this ear." It feels like the ear is plugged, full, underwater. Voices sound wrong. There's often a ringing or hissing. The natural reaction is to suspect wax, a sinus thing, a cold draining funny, or the after-effects of a flight. So people wait a day. Then two. Then they call their primary care office, who may book them next week. Then they try the drugstore wax-removal kit. By the time someone runs a hearing test, the window is closing or gone.
A second trap: the "most people get better on their own" statistic, which is true and dangerous. It means that for two-thirds of people who never do anything, the ear comes back. It also means that for the other third, the ear stays dead forever, and you cannot tell at the start which group you are in. The only way to take the favourable side of that coin is to treat as if you are in the unlucky third Chandrasekhar et al. 2019.
The bedside test that separates "blocked ear" from "dead ear" takes thirty seconds and a tuning fork. Hum, or put a vibrating fork on the middle of your forehead: if the sound goes to your good ear, that's the bad sign — the nerve in the other ear isn't carrying. If it goes to the muffled ear, it's probably just wax or fluid. Most ER doctors and all ENTs know this test. The trouble is getting in front of one.
What happens if you wait
The first week, the odds are with you. Most people who get steroids in the first seven days and have a recoverable loss to begin with get most of their hearing back. By the end of the second week the door is closing; the trial that established the protocol stopped enrolling people at fourteen days for a reason Rauch et al. 2011. Past four to six weeks, the ear is the ear. There are salvage attempts — steroid injections through the eardrum, hyperbaric oxygen — but the returns are smaller and the ceiling is lower Chandrasekhar et al. 2019.
The forecast for a permanent unilateral loss isn't dramatic at first. The first month you adjust: you keep your good ear toward the speaker, you ask people to repeat themselves more often, you stop trying to figure out where a siren is coming from on the street. By the end of the first year you've quietly stopped going to certain restaurants. Group dinners feel like work. Calls on the dead-ear side stop happening; you switch the phone to the other hand without thinking. A persistent ringing in the dead ear keeps you company — tinnitus is the standard package, and there is no off switch.
A decade in, what longitudinal studies of chronic hearing loss find is the social and cognitive cost. People with untreated hearing loss withdraw earlier, get depressed at higher rates, and show faster cognitive decline than peers with intact hearing. The mechanism is partly the load of straining to listen all day; partly the gradual thinning of social contact when conversations get exhausting. The choice the morning you noticed the muffled ear was, looked at honestly, the choice between a 24-hour inconvenience and that decade.
What to do, in order
If one ear has been notably worse than the other for more than a few hours and it isn't hurting, isn't leaking, and isn't obviously just wax, treat it as an emergency. Today, not Monday.
Do not skip the audiogram. Do not accept "it's probably congestion, give it a week" without one. The audiogram is what tells the doctor whether the loss is the nerve (treat now) or the eardrum and middle ear (no rush, often self-resolving).
When the pills are the wrong route
None of these are reasons to skip treatment. They are reasons to pick the other route.
What recovery actually looks like
If you make it in within the first week and the loss is in the recoverable range, the realistic outcome — depending on how bad the initial drop was — is some-to-most of the hearing comes back over the next few weeks, and the ear settles. The ringing fades for many people, though not all. By the end of the steroid course you can usually tell which way it's going.
What you avoid is the slow set of accommodations the "stakes" section walks through. You keep being able to follow a conversation at the end of a noisy restaurant table. You keep being the person friends call rather than the person who's slow to respond. You keep the side of the room you used to ignore. That sounds small — until you talk to anyone who has the other version.
The thing the protocol cannot promise is full recovery for the unlucky cases. Profound loss, especially with vertigo at the start, often comes back only partially regardless of how fast anyone moves. The point of the protocol is not certainty; it is access to the best probability that's available, and a clean conscience about the outcome.
Where this goes wrong in practice
The classic failure is presenting too late. The second classic failure is presenting to the wrong place. Urgent care often does not have an audiometer and the clinician may not run a tuning-fork test confidently — some will refer you to ENT for a week from Tuesday, which is the entire window gone. If the urgent-care clinic can't confirm with audiometry and start steroids the same day, go to the ER instead.
The third failure is the doctor who takes the symptom at face value and treats for an ear infection or wax. If you have been told it's wax and removing the wax hasn't fixed it within a day, that diagnosis was wrong and the clock is still running. Push for an audiogram and an ENT.
The fourth failure is taking the prednisone for three days, feeling better, and stopping. Half a course is half a treatment. Finish the taper.
What this entry doesn't cover
Gradual age-related hearing loss, loss from years of loud noise, hearing changes from ototoxic medications, conductive losses from middle-ear problems, and Meniere's disease all sit next to this topic but follow different timelines and different rules. Chronic tinnitus once the window has closed is a separate body of evidence and a separate set of options. So is unilateral hearing aid use, CROS devices, and cochlear implants for the cases where recovery doesn't come.
- — The reason people miss the window is mistaking this for wax — but wax builds slowly and doesn't strike overnight.
- — Before chasing other causes of a sudden drop, check whether a new medication could be the culprit.
- — Same logic, different sense: sudden loss in one eye or ear is an emergency where hours decide the outcome.
- — A permanent dead ear usually comes with chronic ringing — one more reason to act inside the two-week window.
- — Different problem entirely: this is a sudden, treatable drop, not the subtle in-noise difficulty that tests miss.
- — Age-related loss creeps in over years; this strikes one ear in hours and is a steroid-window emergency.
Substance and claimed effects
Sudden sensorineural hearing loss (SSNHL) is a rapid-onset hearing loss in one ear caused by injury to the cochlea, the auditory nerve, or central auditory pathways. The AAO-HNS clinical practice guideline defines it audiometrically as a sensorineural drop of at least 30 dB across three contiguous frequencies, developing within 72 hours, when compared to the unaffected ear or a prior audiogram Chandrasekhar et al. 2019. In over 90% of cases no cause is ever found (idiopathic SSNHL, or ISSNHL); the remainder are explained by vestibular schwannoma, vascular events, autoimmune inner ear disease, infections (HSV, VZV, CMV, syphilis, Lyme), Meniere's, trauma, demyelination, or ototoxicity. The substance covered here is the condition itself: its recognition, the emergency-window logic that follows from its short steroid efficacy curve, and the trajectory of hearing afterward. Meaningful downstream consequences include residual hearing in the affected ear, chronic tinnitus, hyperacusis, the mood and social-isolation burden of unilateral deafness, and the cognitive-load tax of single-sided hearing in noise.
Evidence by addressing question
Mechanism
Science. The 2019 guideline catalogues the leading mechanistic hypotheses for idiopathic SSNHL: viral cochleitis (latent neurotropic viruses reactivating in the spiral ganglion), microvascular occlusion of the labyrinthine artery (an end-artery with no collateral supply), intracochlear membrane rupture, and autoimmune attack on inner-ear antigens Chandrasekhar et al. 2019. None has been confirmed as the dominant cause; the guideline panel explicitly chose the empirical strategy — treat broadly with corticosteroids because the mechanism is anti-inflammatory enough to cover several plausible pathologies — over waiting for a mechanism-first approach.
Mechanism. The cochlea is uniquely fragile: hair cells do not regenerate in humans, and the stria vascularis depends on a single end-artery. Any inflammatory, ischaemic, or autoimmune insult that outlasts a narrow window kills hair cells permanently. This explains the time-pressure clinically observed even before the mechanism is named: every day of unresolved insult risks structural loss that no later treatment can reverse.
Evidence
Science. Annual incidence in the United States is approximately 27 per 100,000 overall, rising to 77 per 100,000 in adults aged 65+, based on claims-database analysis of more than 60 million enrollees Alexander & Harris 2013. Earlier estimates of 5–20 per 100,000 likely underestimated true incidence because of under-presentation and miscoding as conductive loss NIDCD. The natural history baseline comes from Mattox and Simmons, who reported that roughly 65% of untreated patients recovered to functional hearing levels, most within 14 days Mattox & Simmons 1977. This is the rate any therapy must beat to claim efficacy, and is also why under-treatment so often gets retrospectively excused ("they got better anyway").
The pivotal corticosteroid trial is Rauch et al., a multicentre randomised non-inferiority trial of 250 patients with unilateral SSNHL of ≥50 dB presenting within 14 days. Intratympanic methylprednisolone was non-inferior to oral prednisone at two months; the 95.21% upper confidence bound on the oral-prednisone advantage was 6.6 dB Rauch et al. 2011. The trial established the two-week presentation window as the clinically actionable one and is the basis for both 2012 and 2019 guideline recommendations Stachler et al. 2012 Chandrasekhar et al. 2019.
Recovery is strongly stratified by severity and audiogram shape. Profound losses (>90 dB) almost never recover regardless of treatment; flat and downsloping curves recover poorly (19–56%); low-frequency and upsloping curves recover well (63–88%). Vertigo at onset and age >60 predict worse outcomes.
Practice / clinical consensus. The AAO-HNS guideline grades a strong recommendation that clinicians distinguish sensorineural from conductive loss at presentation, and offer corticosteroids within two weeks of onset as the initial therapy; intratympanic corticosteroids are recommended as salvage within 2–6 weeks Chandrasekhar et al. 2019. Routine head CT, antivirals, thrombolytics, and vasodilators are recommended against. MRI or auditory brainstem response is recommended to rule out retrocochlear pathology (most commonly vestibular schwannoma).
Protocol
Science / Practice. The standard initial regimen drawn from Rauch and codified in the guideline is oral prednisone 1 mg/kg/day (max 60 mg) for 10–14 days, then tapered Rauch et al. 2011 Chandrasekhar et al. 2019. Intratympanic dexamethasone or methylprednisolone, delivered through the eardrum, is offered as initial therapy when oral steroids are contraindicated (poorly controlled diabetes, peptic ulcer disease, psychiatric instability), and as salvage when oral therapy is incomplete at 2–6 weeks. Hyperbaric oxygen therapy is a conditional option ("may be offered") within 2 weeks of onset combined with steroids, or as salvage within one month, with a small absolute hearing-gain benefit reported in meta-analysis but modest study quality Bennett et al. 2012. Audiology referral is recommended for hearing aids, CROS devices, or cochlear implant evaluation when residual loss is unrecoverable.
Contraindications
Practice. The contraindications belong to the treatment, not the diagnosis. Oral high-dose corticosteroids carry risks: hyperglycaemia in diabetes, gastric ulcer in NSAID users, psychiatric destabilisation, avascular necrosis with prolonged courses, immunosuppression. Intratympanic steroids carry small procedural risks (eardrum perforation, infection, dizziness). The guideline panel weighed these against the irreversibility of untreated hearing loss and concluded the benefit-risk balance favours treatment for nearly every otherwise-healthy adult, with intratympanic as the route when systemic steroids are unsafe Chandrasekhar et al. 2019.
Misconceptions
Practice / Community. The dominant clinical-practice failure is misattribution at first contact. Patients themselves, urgent-care clinicians, and primary-care physicians frequently attribute sudden one-sided hearing loss to earwax, a sinus infection, allergies, eustachian tube dysfunction, or a recent cold. The case-by-case error is small (most ear infections do resolve); the systematic error is large, because SSNHL has a steep treatment-window decay and an irrigation, decongestant, or wait-and-see plan burns days off the window. The guideline writes an explicit recommendation against attributing presumptive SSNHL to cerumen or otitis with media before tuning-fork testing Chandrasekhar et al. 2019.
A second misconception is that hearing "always comes back". Mattox and Simmons' 65% spontaneous recovery rate is genuine but conditional — it falls sharply with greater severity, downsloping audiogram, age, and vertigo, and complete spontaneous recovery is rare in severe-to-profound losses Mattox & Simmons 1977.
Stakes
Science. Time-to-treatment is the modifiable variable that most strongly predicts recovery. Across observational series, recovery odds are high in the first week and drop steeply afterward; presentations beyond two weeks lose access to the trial-supported steroid window, and salvage intratympanic steroids have a narrower benefit and a 2–6 week ceiling Chandrasekhar et al. 2019. Permanent unilateral hearing loss compounds: bilateral spatial hearing is lost, speech-in-noise comprehension drops disproportionately, tinnitus often persists in the dead ear, and longitudinal data on chronic hearing loss link it to higher rates of depression, social withdrawal, and accelerated cognitive decline.
Payoff
Science. Patients presenting within the first week and treated with oral prednisone have substantially higher complete-recovery rates than those presenting after 2–4 weeks; the Rauch trial's 14-day enrolment window directly reflects this clinical observation Rauch et al. 2011. The realistic payoff of fast action is not guaranteed full recovery — that depends on initial severity and audiogram shape — but it is the highest achievable probability of full recovery and the cleanest avoidance of permanent unilateral deafness.
Audience
Science. Incidence rises with age, peaking after 65 Alexander & Harris 2013, but cases occur from childhood onward and the largest proportion of new cases is in adults 40–60. Male-to-female ratio is approximately 1:1. Audience scoping for this entry is "all adults" with informational priority for ages 40+ where incidence climbs and pre-existing age-related hearing loss can mask the unilateral drop.
Practicalities
Practice. Emergency department, urgent care, or same-day ENT (otolaryngology) are the appropriate first stops. Audiometry confirms the diagnosis; a Weber tuning-fork test lateralising to the unaffected ear is the bedside screen when audiometry is not immediately available. MRI of the internal auditory canals follows to rule out retrocochlear pathology. Oral steroid courses are inexpensive and widely covered. Intratympanic injection is an in-office otolaryngology procedure typically covered by health insurance. The high-cost downstream interventions are reserved for unrecovered cases: CROS hearing aids ($2–6k), bone-anchored hearing aids ($5–10k), and cochlear implants ($40–100k including surgery and follow-up).
Out of scope
Acoustic trauma, noise-induced hearing loss, age-related sensorineural hearing loss, ototoxic medication injury, conductive losses (otitis, otosclerosis, perforated eardrum), and chronic Meniere's disease all share final-common-pathway features with SSNHL but have separate diagnostic and management pathways and warrant their own entries.
The credibility range
Optimist case. The 2019 AAO-HNS guideline is the strongest evidence-based consensus in otology for a time-sensitive condition. The pivotal Rauch trial is a well-powered multicentre randomised non-inferiority study with an active comparator, not a placebo-controlled trial — but the field considers placebo-controlling a treatable emergency unethical, and the active-comparator design plus the Mattox-Simmons natural-history baseline together give a defensible inferential foundation. Steroid therapy is cheap, widely available, and the time-window evidence is consistent across multiple observational series. Acting fast genuinely changes outcomes.
Skeptic case. The single placebo-controlled steroid trial that exists (Wilson 1980) is small and old, and meta-analyses have not consistently shown steroid superiority over no treatment. About two-thirds of patients recover spontaneously, which means many observed "steroid wins" are misattributions. The 14-day window is a trial enrolment cut-off, not a biologically validated cliff edge; some observational data show benefit at 4–6 weeks too. Hyperbaric oxygen is recommended on the strength of a small effect size from underpowered trials. The guideline panel itself acknowledges the evidence is "C"-grade for much of the protocol.
Author's call. The Skeptic case is technically correct on the trials, but the asymmetric loss function dominates: untreated SSNHL has a real and irreversible failure mode (permanent unilateral deafness), oral corticosteroids are low-cost and low-risk for most otherwise-healthy adults, and the marginal recovery improvement, even if uncertain, comes at trivial cost. The medical-emergency framing is appropriate and the "treat fast" protocol is the right default. Where uncertainty justifies softer language: hyperbaric oxygen sits at the edge of recommended; routine antiviral use is correctly recommended against; the 14-day boundary is a presumption rather than a guarantee — later presentations should still be evaluated and treated, just with lower expected return.
Stakeholder and incentive map
Otolaryngology guideline bodies (AAO-HNS) push the medical-emergency framing because they see the long tail of permanent unilateral deafness cases that arrived too late. Emergency medicine has historically under-recognised the condition because tuning-fork testing is variably taught and the chief complaint "my ear feels blocked" presents non-emergently. Hyperbaric medicine groups push HBOT, with a commercial incentive (HBOT centres charge per session). Hearing-aid and cochlear-implant manufacturers benefit from unrecovered cases but are not the primary advocates for the emergency framing. Patient-advocacy groups (Hearing Loss Association of America) reinforce the emergency framing in lay communication.
Population variability
Best recovery prognoses: younger patients, low-frequency or upsloping audiogram, mild-to-moderate initial loss, no vertigo at onset, no diabetes or cardiovascular comorbidity, presentation within seven days. Worst prognoses: age >60, profound initial loss (>90 dB), flat or downsloping audiogram, accompanying vertigo, hypertension or diabetes, late presentation. Children and elderly patients with profound bilateral loss require modified workup — the bilateral or recurrent presentation specifically demands autoimmune and CNS workup beyond the standard SSNHL pathway Chandrasekhar et al. 2019.
Knowledge gaps
The dominant unknown is mechanism. Without a confirmed cause for the majority of cases, treatment remains empirical and unable to be optimised by pathology. A modern placebo-controlled corticosteroid RCT will not be done because the condition is now considered an emergency and equipoise is gone — the field is locked into the current empirical protocol whether or not steroids exceed the natural-history baseline. The role of antiviral therapy in HSV/VZV-proven cases remains under-studied. The COVID-era association between SARS-CoV-2 and SSNHL incidence is contested; case-control evidence has not consistently shown a population-level effect. Cochlear hair-cell regeneration therapies (atoh1, Notch-pathway, gene therapy) are in early trials and could change the recovery landscape but are not yet clinically available.
Brief vs coverage. The brief named three things: emergency framing, the steroid window, and recovery effects. All three are covered end to end (mechanism + evidence + misconceptions for the emergency framing; protocol + contraindications for the steroid window; stakes + payoff + failure-modes for recovery). Nothing dropped.
Scoring difficulties. The benefit dimensions on a respond condition are inherently indirect — the "benefit" is the downstream loss avoided, not a felt effect of the intervention itself. The pitches were rewritten several times to surface that asymmetry without sounding fear-mongering. energy at 1 and focus at 2 are the smallest defensible non-zero scores; the dossier earns them through the speech-in-noise effort literature, not through a direct cognitive-boost story. longevity at 2 borrows from the hearing-loss / cognitive-decline / dementia association literature; the chain is real but indirect, which is why it sits at 2 rather than 3.
Controversy at 2. The steroid efficacy debate against the natural-history baseline is genuine but mostly a clinician-internal debate at this point; the public-facing field treats the protocol as settled. Scoring 2 reflects that — real disagreement, low temperature.
Excluded on purpose. Long-form post-recovery management (cochlear implants, CROS, bone-anchored hearing aids, chronic tinnitus management) belongs in its own entries — the action-payoff arc here is the emergency response, not the long-tail aftercare. Hearing-loss-and-dementia gets a single referenced sentence in stakes but not a whole section; it warrants its own entry under hearing or mental.
Separate-entry candidates.
- Vestibular schwannoma (acoustic neuroma) — surfaces here as "rule it out with MRI" but is a substantial topic with its own decision tree.
- Chronic tinnitus management — the "there is no off switch" line earns a dedicated entry.
- Hearing-loss-and-cognitive-decline — longitudinal evidence justifies its own entry on the prevention side.
- Cochlear implants and CROS devices for single-sided deafness — the salvage-path entry.
- Hyperbaric oxygen for SSNHL — mentioned in protocol but the controversy and cost-effectiveness merit a focused entry.
Future links. Once written: tuning-fork bedside testing (the 30-second test referenced in misconceptions), audiometry as a screening tool, the hearing-loss / dementia entry.
Tone calibration. The article repeatedly resists the temptation to escalate language — "medical emergency" appears once in the dek and once in the protocol section, deliberately, because the symptom genuinely does present as "probably wax" and over-dramatising loses the audience that needs to hear it. The asymmetric-loss argument carries the urgency instead.
Sudden Hearing Loss
If you act inside the two-week window, steroids can turn a permanently deaf ear into a recovered one.
A generic steroid course and an ER visit. The bill is small compared to what a permanent loss costs you.
Recognise the symptom, go to the ER the same day, take pills for two weeks. Once-only.
A major randomised trial and the standard ENT guideline both back the two-week steroid window.
Persistent one-sided deafness and tinnitus push depression and social withdrawal up. Acting fast keeps that door closed.
Untreated unilateral deafness compounds over decades into depression, isolation, and faster cognitive decline. Fast action prevents that future.
Single-sided hearing costs you mental bandwidth in every meeting and crowded room. Preserving both ears preserves that bandwidth.
Listening with one ear in a noisy room is exhausting. Treating the loss while it's still treatable keeps that daily fatigue from setting in.