A 2025 analysis of 192,993 UK Biobank participants found that just over a fifth reported tinnitus, that hearing health was its main risk factor, and that mood, neuroticism and sleep predicted how severe it was.
UK Biobank is a research resource following the health of about 500,000 people recruited across England, Wales and Scotland between 2006 and 2010, when they were aged 40 to 69, with detailed health, lifestyle and, for many participants, genetic data. Because of its scale, it has become one of the most valuable sources of tinnitus research, able to detect patterns that smaller studies do not have the statistical power to see clearly.
What makes this dataset different
Most tinnitus studies work with hundreds or a few thousand participants. UK Biobank analyses draw on tens or hundreds of thousands, which matters for a condition like tinnitus where individual experience varies enormously — a large sample size is what lets researchers reliably separate real patterns (age, sex, specific risk factors) from statistical noise.
Key findings from the risk-factor analysis
The 2025 analysis, published in Nature Communications, used machine learning to predict two things separately from socio-demographic, psychological and health information: whether a participant had tinnitus, and how severe it was. Of the 192,993 participants, 41,042 had tinnitus.
- Overall prevalence: just over a fifth of participants reported tinnitus, with roughly a quarter of those describing it as moderately or severely distressing — a higher figure than some national estimates, likely reflecting differences in survey methodology and population rather than a higher underlying rate in the UK. Tinnitus statistics explains why those estimates differ.
- Risk factors: hearing health was the primary risk factor for both having tinnitus and its severity. Mood, neuroticism and sleep predicted severity; sleeping with tinnitus covers the sleep side of that.
- Change over time: the severity model predicted how participants' tinnitus progressed over nine years, including a large effect for those who went on to develop severe tinnitus.
- Age pattern: prevalence rose steadily from around age 40, peaked in the 65–79 range, and then declined somewhat in the oldest participants — a pattern consistent with other tinnitus research. Because the cohort was recruited at ages 40 to 69, it says little about tinnitus in younger adults.
- Sex differences: men reported higher overall prevalence consistently from ages 40 through 79, while women in the study tended to report greater distress when tinnitus was present — suggesting sex differences in tinnitus aren't simply about who gets it, but potentially about how it's experienced once present.
What the genetic side of UK Biobank has added
Because many UK Biobank participants have genetic data, the resource has powered some of the most detailed tinnitus genetics research to date, covered in more depth in genetics and tinnitus.
A 2020 genome-wide association study of 172,995 UK Biobank participants found tinnitus to be polygenic — shaped by many genes of small effect — identified 6 risk loci, and found a genetic link with major depressive disorder. In 2024 the same research group combined UK Biobank with the US Million Veteran Program, 596,905 people in all, and identified 39 risk loci. That larger study found tinnitus shares many genetic variants with hearing difficulty but also has a distinct genetic architecture of its own, and it confirmed genetic correlations with psychiatric conditions, depression among them.
Why this kind of research matters for patients, even though it doesn't change today's treatment
Large-scale studies like this one aren't designed to produce an immediate new treatment — their value is in building an accurate, reliable picture of who's affected, how, and why, which shapes research priorities and eventually informs more targeted treatments. The genetic and statistical links between tinnitus and depression found in this dataset, for instance, add weight to the case for treating the two together rather than assuming they're unrelated, which tinnitus and depression covers.
A caution worth keeping in mind
UK Biobank participants are not representative of the general population. Only 5.5% of the roughly 9.2 million people invited took part, and compared with the general population, participants were less likely to be obese, to smoke or to drink alcohol daily, had fewer self-reported health conditions, and lived in less socioeconomically deprived areas — a "healthy volunteer" effect. The researchers who documented this concluded that relationships between exposures and disease can still be generalised, even though the cohort's percentages may not describe the country. That is a good way to read the tinnitus findings: the risk-factor relationships are more transferable than the exact prevalence figure.