The biggest test of the hearing loss dementia link gave hearing aids to 977 older adults and followed them for three years. Thinking and memory slipped at the same rate in both groups (Trial). Decline did run about half as fast in one higher-risk quarter of the group. That is one slice of the trial. Hearing loss still tops the Lancet Commission’s list of dementia risks you can change (Report). Treating it is worth doing either way.
What Hearing Loss Actually Is
Age-related hearing loss has a name of its own: presbycusis. It takes the high notes first, and consonants live up there, along with children’s voices and birdsong. Volume is the last thing to go. That is why the same sentence comes up again and again: I can hear you, I just cannot understand you. In a quiet room people manage fine. In a restaurant they are lost.
This is the normal state of an older ear rather than a rare condition. One national US study measured hearing in 2,803 adults aged 71 and over (Study). About 65 in every 100 had some hearing loss, which works out at roughly 21.5 million people. The figure ran from 53 in 100 at ages 71 to 74, up to 96 in 100 past 90.
Far fewer of them do anything about it. Among those with measured hearing loss, fewer than 3 in 10 wore a hearing aid (Study). The US National Institute on Deafness puts the number of American adults who could benefit from a hearing aid at about 28.8 million (NIDCD).
Hearing fades so slowly that people adapt without noticing. They turn the television up, skip the noisy pub and work around it.
How Hearing Loss Reaches the Brain
Researchers have three explanations for how a worn-out ear might affect thinking. Each one is plausible, and none of them is settled.
The first is simple lack of input. In a Baltimore ageing study, 126 adults had repeat brain scans over about six years (Study). Those with hearing loss lost brain volume a little faster, about 8.4 cubic centimetres a year against 7.2. The extra loss sat in the right temporal lobe, the part that handles sound and speech.
Another study points the other way. A 2026 Australian study scanned 312 older adults twice, three years apart (Study). Neither hearing loss nor hearing aid use was linked to faster brain change or to thinking scores. What did track faster thinning was trouble following speech in noise, and that points more at the brain than at the ear.
The second idea is effort. Straining to decode half-heard words uses attention that could go to holding the sentence in mind. No trial has tested it.
The third is withdrawal, and it is the only one of the three with trial evidence behind it. The hearing aid trial also tracked social life (Trial). Over three years, the hearing group held on to about one more person in their social network than the control group. That was not what the trial set out to measure, and the margin was slim. Loneliness scores moved slightly their way too, though the authors say that change is too small to matter in daily life (Trial). A shrinking social circle carries its own health cost, which is why loneliness as a health risk has its own article here.
Which of the three routes matters most is still an open question. So is whether restoring sound reverses the brain change. The trial ran brain scans too. So far those results exist only as conference abstracts, and the effect was smallest in the hearing part of the brain (Conference abstract).
The Trial That Tested Hearing Aids
Most dementia prevention advice comes from watching people over time. Hearing loss is one of the few risk factors anyone has actually tested in a trial.
The trial is called ACHIEVE. It ran at four US sites and was paid for by the National Institutes of Health (Trial). It enrolled 977 adults aged 70 to 84 who had untreated hearing loss and no serious memory problems. A coin flip decided who got what.
490 people received hearing aids plus counselling from an audiologist. The other 487 got sessions with a health educator about ageing well. Both groups got something, so the hearing aids were compared against a real alternative.
The researchers followed everyone for three years. They measured thinking and memory with a full set of tests rather than one short quiz. Everyone counted towards the group they were assigned to, even the people who stopped wearing the aids.
Both groups declined by the same amount (Trial). The authors’ own summary opens with that result. Hearing aids did not slow the loss of thinking and memory across the trial as a whole. The study reported no serious harms.
A trial that finds nothing is still useful, because it tells you the effect is not large in the people it enrolled. Most news coverage skipped it and reported a number from one slice of the trial instead.
Why the Subgroup Result Is Weaker
The headline everyone read came from 238 of those 977 people.
The trial recruited from two streams. 739 answered adverts, and as a group they were healthy. The other 238 came from ARIC, a long-running US heart-health study. They were about three years older on average, with less education, more diabetes and more high blood pressure. They more often lived alone, and they started with lower test scores.
In that group, decline ran about half as fast with hearing aids (Trial). Among the healthy volunteers there was no benefit. The trial itself files that result as a planned extra check on where people were recruited, rather than as its answer.
A trial can be sliced many ways. Cut it ten times by the usual standards and the odds of one slice looking real by accident are better than 4 in 10 (Review). One trial described in that review planned 60 separate subgroup comparisons.
The same pattern shows up across the research. In one audit, 207 trials reported subgroup analyses and 64 claimed an effect on the main result (Systematic review). Judged against ten credibility checks, 54 of those claims passed four or fewer. A second audit found 46 subgroup claims with solid statistics behind them (Analysis). Only 5 were ever retested in a later trial or combined analysis, and none of the 5 survived.
Size matters too. 238 people is a quarter of the trial. The healthy volunteers barely declined at all over three years, so there was very little decline left to slow.
A subgroup result is a good reason to run another trial. It is a weak reason to believe the number.
Researchers later re-sorted the same 977 people by predicted risk instead of by recruitment source (Trial). The highest-risk quarter again declined roughly 60% more slowly with hearing aids, and the rest showed nothing. It is the same people, cut a different way, and the answer came out the same. That is encouraging. What would settle it is a fresh trial that enrols only higher-risk older adults.
Where Hearing Loss Dementia Risk Ranks
Line up everything you can change about dementia risk, and hearing loss comes first.
The 2024 Lancet standing Commission lists 14 risk factors you can act on (Report). Together they are linked to about 45% of dementia cases worldwide. Hearing loss carries the largest single share at about 7%, roughly level with high LDL cholesterol. PM2.5 air pollution sits on the same list.
That 7% rests on some big assumptions. It assumes hearing loss really does cause dementia. It also assumes you could remove hearing loss from the population completely. On top of that, the Commission strips out the overlap between risk factors using one Norwegian study of 37,000 people (Report). Before that adjustment, the share for hearing loss was about 19%.
The risk estimate underneath is thinner than the ranking suggests. The Commission pooled six studies covering 666,370 people (Meta-analysis). People with measured hearing loss were about 37% more likely to develop dementia. That figure is shaky. The studies disagreed with each other a lot, and the evidence still leaves room for no extra risk at all. Still, every one of the four studies that checked found risk climbing as hearing got worse.
Then there is the direction problem. Early brain change might cause the hearing trouble rather than the other way round. Genetic studies are the usual way to test that, and they do not support cause here. A systematic review graded every hearing loss analysis as insufficient evidence (Systematic review). Insufficient is not the same as disproven, and there were only a handful of analyses to grade. Another study found almost no shared genetics, and the one signal it did find ran backwards, from Alzheimer’s risk genes to hearing difficulty (Study).
So the Commission’s ranking and the flat trial result do not really conflict. One is a population estimate built on a best case. The other is what three years of hearing aids did in 977 people.
People who already wear aids have been studied too. Pooled data across 126,903 people found roughly a fifth fewer cases of decline or dementia among device users (Meta-analysis). Every study in that pool only watched people, and the authors rated the evidence as weak. Those users also differ from everyone else. In a UK study of 59,768 adults, aid users at first looked more likely to develop dementia (Study). Adjusting for how often people saw a doctor flipped that finding, and left it small enough that chance could explain it.
Hearing Aids You Can Buy Today
In August 2022 the US changed the rules (NIDCD). Hearing aids for mild to moderate loss can now be sold in shops and online. No medical exam is needed, and no fitting by an audiologist. The category covers adults aged 18 and over. Devices for severe loss, and for anyone under 18, still need a prescription.
Self-fitting works for the right people, at least in one small trial. A coin flip split 68 adults with mild to moderate hearing loss between two routes (Trial). One group fitted the same device themselves with remote support. The other had an audiologist fit it using best practice. At six weeks both groups reported the same benefit, and both did the same on understanding speech in noise. A follow-up at about eight months found them still level (Trial). Only 44 people came back for it, and satisfaction slipped in both groups.
That trial had real limits. It used one device, at one clinic, and everyone knew which group they were in. The company that makes the device helped pay for the work. Over-the-counter aids are not meant for severe loss or for children.
Some symptoms mean seeing a clinician rather than buying a device (NIDCD). Ear pain is one. Fluid, blood or pus from the ear in the past six months is another. Get checked if your hearing drops suddenly, keeps getting worse, or comes and goes. Dizziness alongside the hearing loss needs a look. So does loss or ringing in one ear only, or a clear difference between your ears.
Our best national counts of who treats their hearing loss were taken before the shelves changed. So nobody can yet say how far the gap has closed.
Either way, start with a real hearing test. It is cheap or free in most places, and it tells you which route you are on.
Protect the Hearing You Have
Hearing damage adds up, and it does not grow back. Protecting your ears now is cheaper than treating them later. Most of that exposure is now voluntary.
Researchers pooled 33 studies of people aged 12 to 34 (Meta-analysis). About 1 in 4 listened to personal devices at unsafe levels. An estimated half were exposed to unsafe sound at clubs, gigs and bars. The WHO puts it more bluntly: over a billion young adults risk permanent, avoidable hearing loss (WHO).
There is a safe budget for sound. It is about 80 decibels averaged across 40 hours a week (WHO). That budget shrinks fast as things get louder: you get 4 hours at 90 decibels and 20 minutes at 100. A rough rule for headphones is to stay at or below 60% of maximum volume.
Four habits cover most of it. Turn the volume down, and use noise-cancelling headphones instead of turning it up. Wear plugs at gigs and with power tools. Get a baseline test while your hearing is still good.
Frequently Asked Questions
Is the hearing loss dementia link cause or coincidence?
Nobody can say for certain yet. Studies that simply follow people find a large and consistent link. Pooled across 666,370 people, hearing loss was linked to roughly 37% higher dementia risk (Meta-analysis). Genetic studies are designed to test cause, and they do not back it up. One systematic review graded every hearing loss analysis as insufficient evidence (Systematic review).
Do hearing aids prevent dementia?
The one proper trial says no, at least over three years in the people it enrolled (Trial). Studies that only watch people give a friendlier answer. Across 61,089 adults in 33 countries, hearing aid users had about 9% lower dementia risk (Study). That benefit disappeared in people who said the aid had not improved their hearing. So wearing one may help, though it is a long way from proven protection.
Are over-the-counter hearing aids as good as prescription ones?
For mild to moderate loss, a small trial found no meaningful difference (Trial). People who fitted the device themselves with remote support did as well at six weeks as those fitted by an audiologist. The two groups were still level at about eight months (Trial). The category is limited to adults with mild to moderate loss (NIDCD). Severe loss still needs a professional.
At what age should I get a hearing test?
Earlier than most people do. Disabling hearing loss affects about 22 in every 100 US adults aged 65 to 74, and 55 in every 100 past 75 (NIDCD). By age 71 and over, about 65 in 100 have some measurable loss (Study). Sudden loss, one-sided loss, pain or discharge means seeing a clinician straight away, at any age.
Does treating hearing loss help if I already have memory problems?
The trial cannot answer that. It deliberately excluded people with substantial memory impairment (Trial). The group that appeared to benefit was at higher risk of decline, and not already impaired (Trial). If memory is already a worry, that is a conversation to have with a clinician.
Key Takeaways
- The whole trial found nothing. 977 older adults got hearing aids or health education for three years, and both groups declined by the same amount (Trial).
- The famous “about half” came from a slice. It describes 238 higher-risk participants, which the trial files as a planned extra check (Trial). Subgroup claims rarely survive checking: of five ever retested, none held up (Analysis).
- Hearing loss still ranks first among changeable risks. The 2024 Lancet Commission gives it the largest single share of 14 factors (Report). That number assumes cause, and assumes hearing loss could be removed completely.
- The social benefit came out of the trial itself. Treated participants kept about one more person in their social network over three years (Trial).
- Buying without a prescription is a legitimate route. Self-fitted over-the-counter devices matched audiologist-fitted ones at six weeks and at eight months (Trial, Trial).
- Noise damage is permanent and preventable. About 1 in 4 young people listen unsafely, and the safe budget is roughly 80 decibels over 40 hours a week (Meta-analysis, WHO).
Start With a Hearing Test
The action here is small, and it does not depend on the dementia question being settled.
Book a hearing test. If it shows loss, treat it now rather than at 75. Then protect what you have: turn the volume down, wear plugs at gigs, and reach for noise-cancelling headphones instead of more volume.
The rest of the case stands on its own. Hearing better means conversation stops being work, and meetings and meals with family get easier. The trial did show one thing, measured and decided by coin flip: treating hearing loss helped people keep their social circle over three years (Trial).
The dementia evidence is still building. The rest of the reasons to treat your hearing do not need to wait for it.
This article is for educational purposes and is not medical advice. Talk to a qualified clinician before changing your health regimen.
Leave a comment