Two of the fourteen modifiable risk factors on the Lancet Commission's list were added in the 2024 update: raised LDL cholesterol in midlife, and untreated vision loss in later life. Vision is the one almost nobody has heard of, and where the action is cheapest and least interesting. An eye test.
The word doing the work is "untreated"
The factor is not vision loss. It is vision loss that nobody corrected.
That is the whole reason it sits on a list of modifiable risks. Refractive error, cataract and much diabetic eye disease are correctable, and a risk factor you can remove with glasses or a day-case operation is a different proposition from one you can only monitor. The Commission's recommendation follows: eye tests available to everyone, alongside hearing aids.
The same construction appears for hearing, and we have covered that factor separately. The two get bundled as "sensory impairment"; the parallel is real, the evidence bases are not.
What the cohort evidence shows
The largest pooled estimate comes from Shang and colleagues in Ophthalmology in 2021: across 14 prospective cohorts covering 6,204,827 participants and 171,888 people who developed dementia, visual impairment carried a pooled relative risk of 1.47 (95% CI 1.36–1.60). The authors call the overall quality of evidence low: every contributing study was observational.
A 2024 umbrella review in Ageing Research Reviews graded the whole literature (13 meta-analyses, 232 articles, some 99 million participants) against formal credibility criteria, and nothing reached "highly suggestive" or "convincing" evidence. Visual impairment and dementia came out at an equivalent odds ratio of 1.50 (95% CI 1.23–1.84), with I² = 45%, a prediction interval of 1.09 to 2.08 and detectable small-study effects — graded "weak".
In UK Biobank, visual impairment was associated with all-cause dementia at HR 1.55 (95% CI 1.18–2.04). The Mendelian randomisation analysis in the same paper — genetic variants used to get nearer a causal test — did not find what the cohort data predicts: genetically proxied vision impairment was associated with non-Alzheimer, non-vascular dementia (OR 1.62, 95% CI 1.13–2.33), but not with all-cause dementia or Alzheimer's. Hearing impairment, in the same analysis, was.
Cataract surgery: the strongest signal, and its ceiling
The most cited evidence analysed the Adult Changes in Thought cohort in JAMA Internal Medicine, 2022: 3,038 Kaiser Permanente Washington members aged 65 and over, dementia-free at enrolment and with a cataract or glaucoma diagnosis, followed across 23,554 person-years, during which 853 developed dementia. Cataract extraction was associated with a hazard ratio of 0.71 (95% CI 0.62–0.83, P<0.001) for all-cause dementia, and similar for Alzheimer's dementia.
What lifts this above a plain association is the negative control. They also examined glaucoma surgery — comparable seriousness and population, no restoration of vision. If the cataract finding merely marked being well enough for surgery and well looked after, glaucoma surgery should show the same benefit. It did not: HR 1.08 (95% CI 0.75–1.56, P=0.68). They also weighted for surgery probability using marginal structural models, against healthy-patient bias.
It remains observational, and the authors say so: the results "could be explained by unmeasured or residual confounding, like any observational study." People who get cataract surgery differ in ways records capture imperfectly — frailty, anaesthetic fitness, healthcare access, somebody's judgement that the cataract warranted operating on. The cohort was also 91% self-reported White and insured.
A 2024 systematic review in Ophthalmology (24 articles, 558,276 participants) pooled the association at HR 0.75 (95% CI 0.72–0.78) against uncorrected cataract, with GRADE certainty very low to low, and found operated patients indistinguishable from people who never had cataracts (HR 0.84, 95% CI 0.66–1.06). Those authors call for randomised trials.
Where vision and hearing part company
Hearing has a large randomised trial. ACHIEVE randomised 977 adults aged 70 to 84 to hearing intervention or an active control for three years; its primary outcome was flatly null (p=0.96), with a positive result confined to the higher-risk ARIC subgroup of 238 people.
Vision has no equivalent. No large trial has randomised vision correction with dementia or cognitive decline as its primary endpoint, so this literature has never faced the test that flattened hearing's headline claim. It looks cleaner partly because it has been examined less severely.
The competing explanations, and only one is good news
Shared pathology. The retina is developmental brain tissue. Retinal amyloid-β deposits, abnormal tau isoforms, reduced retinal blood flow and blood-retinal barrier damage have all been described in mild cognitive impairment and Alzheimer's. In UK Biobank, among 32,038 people screened to exclude eye disease, diabetes and vision loss, those in the two thinnest quintiles of retinal nerve fibre layer were about twice as likely to do worse on repeat cognitive testing three years later (OR 1.92, 95% CI 1.29–2.85; OR 2.08, 95% CI 1.40–3.08). On that reading poor vision is partly a window onto the brain, not a cause — and the small-vessel disease behind much of it is the cluster blood pressure control and stroke prevention already address.
Reduced input and withdrawal. Reading, faces and navigation are among the most cognitively demanding things most people do, and degraded vision makes them harder or abandoned — the cognitive reserve argument, still a hypothesis. People who cannot see well also stop driving and go out less, and social isolation and inactivity are themselves on the list, so vision loss may act partly through other factors.
Detection bias. Most cognitive tests require seeing something; fail to read the card and you score worse without being more impaired. The umbrella review shows the shape of it — impairment identified cross-sectionally was associated with cognitive impairment at eOR 2.37, against 1.56 when measured objectively. A 2024 analysis in Innovation in Aging tested this directly in LASI-DAD and found the bias small relative to the tests' standard error: reassuring, but one study in one population.
Does it matter which eye disease?
Yes, and not in the direction the mechanism stories predict. From the umbrella review:
- Cataract — dementia eOR 1.20 (95% CI 1.16–1.25), graded suggestive, the highest grade awarded.
- Diabetic retinopathy — dementia eOR 1.33 (95% CI 1.17–1.50), graded weak. Diabetes is separately on the list, and the metabolic route to cognitive risk is hard to separate from the ocular one.
- Age-related macular degeneration — Alzheimer's eOR 1.27, graded suggestive (interval reported as 1.27–1.27 — treat that precision with suspicion). A meta-analysis of 8 studies (8,223,581 participants) put AMD's association with subsequent dementia at pooled HR 1.22 (95% CI 1.01–1.47) — a lower bound only just clear of 1.0.
- Glaucoma — dementia eOR 1.01 (95% CI 0.87–1.18). No association. A 2021 meta-analysis of 11 cohorts (4,645,925 participants) agreed, for Alzheimer's (RR 1.03, 95% CI 0.93–1.05) and all-cause dementia (RR 1.08, 95% CI 0.97–1.19), with I² above 79%; a 2024 meta-analysis of 27 studies found a positive overall association but none in Europe or North America.
Glaucoma is the instructive null: a neurodegeneration of retinal ganglion cells with the strongest mechanistic claim to a shared pathway and the weakest link to dementia, while cataract — purely optical, fully reversible — has the strongest. That fits the correctable-input story better than the common-cause story — the one hopeful thing here. We do not print the Commission's per-factor share: published summaries disagree and the factors overlap.
What to actually do
Have your sight tested on a schedule, not when something feels wrong. Glaucoma is the reason: the NHS says plainly that it "does not usually have symptoms and most people do not realise they have it" and is "usually picked up during routine eye tests". AMD is sometimes found on a routine test before symptoms appear, and because it often starts in one eye the other can mask it. NHS advice is a test at least every two years, or more often if advised.
England, Wales and Northern Ireland. Free NHS sight tests for the under-16s, 16–18s in full-time education, the over-60s, anyone registered blind or partially sighted, anyone diagnosed with diabetes or glaucoma, anyone 40 or over with a parent, sibling or child who has glaucoma, anyone told by an ophthalmologist they are at glaucoma risk, and on qualifying low-income grounds.
Scotland. NHS eye examinations are free to everyone living in the UK — every two years, or yearly if you are under 16, 60 or over, or diabetic.
Ireland. The Treatment Benefit Scheme gives a free eyesight test once every second calendar year if your PRSI record qualifies; tests for visual display units and driving licences are excluded. Medical card holders get a free HSE eye examination every two years, more often if a medical condition requires. Everyone with type 1 or type 2 diabetes aged 12 and over is eligible for free Diabetic RetinaScreen, usually annually.
If your real worry is your memory rather than your eyes, that is a different article; the fuller factor-by-factor account sits here.
What we are not going to sell you
There is no supplement on the Commission's list of fourteen, and we sell nothing for eye health. The AREDS and AREDS2 formulations do have real trial evidence, for a narrow purpose — slowing progression to advanced AMD in people who already have specific intermediate disease. That is a clinical intervention for a diagnosed condition, not general eye support.
And here is the part that would flatter us if written loosely. AREDS2 randomised 4,203 people and tested whether adding DHA 350mg plus EPA 650mg to the AREDS formulation helped. It did not. Progression to advanced AMD by five years was 31% on placebo and 31% on DHA+EPA — hazard ratio 0.97 (98.7% CI 0.82–1.16), P=0.70. Our omega-3 provides 300mg of EPA and DHA combined, well under a third of the dose that produced no benefit in the trial built to test it. There is a serious literature on omega-3 and cognitive ageing worth reading on its own terms; it is not an eye-health claim and we will not borrow one.
It is an optometrist, not a bottle.
This article is information, not medical advice. It cannot tell you whether you have an eye condition or whether any treatment, cataract surgery included, is right for you. Entitlements can change.
Sources: Livingston G, Huntley J, Liu KY, et al. Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission. Lancet. 2024;404(10452):572–628. PubMed · Shang X, Zhu Z, Wang W, Ha J, He M. The Association between Vision Impairment and Incidence of Dementia and Cognitive Impairment: A Systematic Review and Meta-analysis. Ophthalmology. 2021;128(8):1135–1149. doi:10.1016/j.ophtha.2020.12.029 · Rahmati M, Smith L, Lee H, et al. Associations between vision impairment and eye diseases with dementia, dementia subtypes and cognitive impairment: An umbrella review. Ageing Research Reviews. 2024;101:102523. doi:10.1016/j.arr.2024.102523 · Lee CS, Gibbons LE, Lee AY, et al. Association Between Cataract Extraction and Development of Dementia. JAMA Internal Medicine. 2022;182(2):134–141. doi:10.1001/jamainternmed.2021.6990 · Yeo BSY, Ong RYX, Ganasekar P, Tan BKJ, Seow DCC, Tsai ASH. Cataract Surgery and Cognitive Benefits in the Older Person: A Systematic Review and Meta-analysis. Ophthalmology. 2024;131(8):975–984. doi:10.1016/j.ophtha.2024.02.003 · Jiang F, Dong Q, Wu S, et al. A comprehensive evaluation on the associations between hearing and vision impairments and risk of all-cause and cause-specific dementia: results from cohort study, meta-analysis and Mendelian randomization study. BMC Medicine. 2024;22:518. doi:10.1186/s12916-024-03748-7 · Ko F, Muthy ZA, Gallacher J, et al. Association of Retinal Nerve Fiber Layer Thinning With Current and Future Cognitive Decline. JAMA Neurology. 2018;75(10):1198–1205. doi:10.1001/jamaneurol.2018.1578 · Gaire BP, Koronyo Y, Fuchs DT, et al. Alzheimer's disease pathophysiology in the Retina. Progress in Retinal and Eye Research. 2024;101:101273. doi:10.1016/j.preteyeres.2024.101273 · Ehrlich JR, Nichols E, Chen Y, et al. The Influence of Vision Impairment on the Measurement of Cognition in Older Adults in India: Findings From LASI-DAD. Innovation in Aging. 2024;8(9):igae071. doi:10.1093/geroni/igae071 · Tsai HR, Lo RY, Liang KH, et al. Risk of Subsequent Dementia or Alzheimer Disease Among Patients With Age-Related Macular Degeneration: A Systematic Review and Meta-analysis. American Journal of Ophthalmology. 2023;247:161–169. doi:10.1016/j.ajo.2022.11.005 · Zhao W, Lv X, Wu G, et al. Glaucoma Is Not Associated With Alzheimer's Disease or Dementia: A Meta-Analysis of Cohort Studies. Frontiers in Medicine. 2021;8:688551. doi:10.3389/fmed.2021.688551 · Wang X, Chen W, Zhao W, Miao M. Risk of glaucoma to subsequent dementia or cognitive impairment: a systematic review and meta-analysis. Aging Clinical and Experimental Research. 2024;36:172. doi:10.1007/s40520-024-02811-w · Age-Related Eye Disease Study 2 Research Group. Lutein + zeaxanthin and omega-3 fatty acids for age-related macular degeneration: the Age-Related Eye Disease Study 2 (AREDS2) randomized clinical trial. JAMA. 2013;309(19):2005–2015. doi:10.1001/jama.2013.4997 · NHS — Glaucoma · NHS — Free NHS eye tests and optical vouchers · gov.ie — Treatment Benefit Scheme · Citizens Information — Sight tests and eye health · HSE — Diabetic RetinaScreen