Skip to main content
NCBI home page
Alzheimer's & Dementia : Diagnosis, Assessment & Disease Monitoring logoLink to Alzheimer's & Dementia : Diagnosis, Assessment & Disease Monitoring
. 2026 Sep 25;18(3):e70499. doi: 10.1002/dad2.70499

Severity and progression of visual impairment and risk of dementia

Allen T C Lee 1,✉, Owen N W Leung 1, Zhaohua Huo 1, Jessie O T Kwok 1, Kelvin K L Chong 2, Wai C Chan 3, Ada W T Fung 4, Suk L Ma 5, Ruby S Y Lee 6, Marcus Richards 7, Sheung‐Tak Cheng 8, Linda C W Lam 1
PMCID: PMC13615333  PMID: 42801054

Abstract

INTRODUCTION

Visual impairment is associated with dementia risk, but the role of severity and longitudinal change remains unclear. We examined whether visual impairment severity and its progression were associated with incident dementia.

METHODS

This territory‐wide cohort study followed 15,576 community‐dwelling older adults without dementia at baseline for 16 years. Incident dementia was ascertained using International Classification of Diseases, Tenth Revision diagnoses or Clinical Dementia Rating scores of 1 to 3. Visual acuity was measured using the Snellen chart. Cox models estimated hazard ratios, excluding dementia cases occurring within the first 6 years to minimize reverse causation and adjusting for demographic, comorbidity, and lifestyle factors.

RESULTS

Moderate to severe visual impairment and faster visual acuity decline were associated with increased dementia risk. Severe visual impairment was associated with dementia onset 2.1 years earlier than normal vision.

DISCUSSION

These findings strengthen the evidence for visual impairment as a dementia risk factor, supporting incorporation of visual assessment into dementia risk stratification.

Trial Registration

The Chinese Clinical Trial Registry: ChiCTR1900023536

Keywords: dementia, neurocognitive disorder, poor visual acuity, risk reduction, visual impairment

Highlights

  • Poorer visual acuity was independently associated with higher dementia risk.

  • Moderate to severe visual impairment linked to earlier dementia.

  • Rapid visual acuity decline further increased dementia risk.

  • Improving visual acuity lowered dementia risk compared to continued decline.

  • Assessing visual severity and progression may improve dementia risk stratification.

1. BACKGROUND

Visual impairment, which is prevalent in late life yet often underdiagnosed and undertreated, 1 , 2 , 3 is increasingly recognized as a potentially modifiable risk factor for dementia. 4 Although systematic reviews and meta‐analyses have supported an association between visual impairment and dementia risk, 5 , 6 , 7 , 8 several important gaps remain. First, most prior studies were cross‐sectional or retrospective in design and treated visual impairment as a binary exposure, limiting understanding of potential dose–response relationships and whether dementia risk differs between mild or stable visual impairment and severe or progressive visual loss. Clarifying the roles of severity and progression of visual impairment in relation to dementia risk is of great clinical and public health importance, particularly for informing targeted screening and intervention strategies. Second, visual impairment may compromise performance on vision‑dependent cognitive testing, increasing the risk of misclassifying visual deficits as dementia. Although vision‑adapted cognitive screening tools exist, 9 , 10 they do not fully account for confounding factors such as depression. These limitations underscore the importance of comprehensive clinical assessments to accurately distinguish dementia from visual impairment and minimize misclassification. Third, dementia has a long preclinical phase – often lasting a decade or more – during which neuropathological changes precede clinical symptoms. For instance, visual disturbances and retinal nerve fiber layer thinning may occur before cognitive decline becomes apparent. 11 , 12 Consequently, associations observed close to the time of dementia diagnosis may reflect reverse causation rather than a true antecedent relationship. To address this concern, the Lancet Commission on Dementia Prevention recommends excluding dementia cases occurring within the first 5 years of follow‐up in analyses of putative dementia risk factors. 4 Long‐term cohort studies incorporating this approach are therefore essential to more clearly delineate the temporal relationship between visual impairment and dementia.

RESEARCH IN CONTEXT

  1. Systematic review: We reviewed the available literature on PubMed, the Cochrane Database of Systematic Reviews, and Google Scholar from the inception of the databases to February 28, 2026, to identify visual impairment as a risk factor for dementia. The following searched terms were used, and the search was limited to studies published in English: “aged” or “older adults,” “visual impairment” or “poor vision,” “dementia,” and “neurocognitive disorders.” Evidence from the retrieved studies showed that visual impairment might contribute to the development of dementia. However, the influence of its severity and progression on dementia onset is not well understood.

  2. Interpretation: This territory‐wide cohort study followed 15,576 older adults for 16 years to examine whether visual impairment was independently associated with dementia risk and how its severity and progression contributed to that risk. We found poorer visual acuity was associated with increased dementia risk, with moderate to severe impairment and rapid decline conferring the highest risk. Individuals with severe visual impairment developed dementia approximately 2.1 years earlier than those with normal vision. Improvements in visual acuity appeared to attenuate this risk, suggesting that even late‐life interventions may have protective effects, though this warrants confirmation through long‐term trials.

  3. Future directions: Our findings reinforce visual impairment as an independent and modifiable risk factor for dementia. The results highlight the importance of incorporating visual assessments with longitudinal monitoring into dementia risk stratification.

By combining prospective longitudinal assessments of visual acuity with clinical ascertainment of dementia and exclusion of early incident dementia cases, this study aimed to clarify whether the severity and progression of visual impairment were independently associated with long‐term dementia risk while minimizing potential biases from reverse causation, confounding, and outcome misclassification. The findings may strengthen the evidence for visual impairment as a risk factor for dementia and highlight the need to incorporate longitudinal visual assessments into dementia risk stratification.

2. METHODS

2.1. Study design, setting, and participants

This study was based on our longitudinal, territory‐wide community cohort of Chinese adults aged 65 years or older in Hong Kong. 13 The cohort comprised 18,298 individuals who attended any of the 18 Elderly Health Centers (EHCs) – government‐run centers providing primary health care for older adults through health assessments and disease prevention – between January 1 and June 30, 2005. For the present study, individuals with clinically diagnosed dementia at baseline were excluded. To minimize residual confounding from suspected dementia, those living in care homes or with a diagnosis of Parkinson's disease or stroke were also excluded. Additionally, those with incomplete baseline data were excluded. After exclusions, 15,576 participants were included in this study (Figure 1).

FIGURE 1.

FIGURE 1

Open in a new tab

Flow diagram of inclusion and follow‐up of participants in this 16‐year follow‐up study. EHCs, Elderly Health Centers.

Participants were followed for 16 years. Standardized clinical assessments of cognitive status, together with comprehensive evaluations of lifestyle behaviors, physical and psychiatric comorbidities, and demographic characteristics, were conducted at baseline and during regular follow‐ups at the EHCs. To minimize loss to follow‐up, participants who missed scheduled visits were actively traced and assessed by geriatric psychiatrists from the research team. For participants who could not be traced, their names were cross‐checked with the Death Registry to identify dementia recorded as cause of death. Written informed consent was obtained from all participants or from their family members or legal guardians if they became mentally incapable of giving consent before the follow‐up assessment. Ethics approval was granted by the Ethics Committees of the Department of Health of the Government of the Hong Kong Special Administrative Region (SAR) and the Chinese University of Hong Kong. This study was performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.

2.2. Assessment of visual acuity

Habitual visual acuity was assessed using a standard 6‐m Snellen E chart. Participants were tested while wearing their usual optical corrections. Snellen acuity was converted to LogMAR units for analysis, with higher values indicating poorer visual acuity.

2.3. Ascertainment of dementia

The study outcome was incident dementia over 16 years of follow‐up. Dementia ascertainment was based on standardized clinical assessments conducted uniformly by trained physicians at baseline and follow‐up, including detailed history‐taking and non‐visual cognitive tests (delayed recall and the Abbreviated Mental Test 14 ) to minimize misclassification related to vision‐dependent testing. Participants who were traced and interviewed by geriatric psychiatrists underwent additional clinical evaluation and assessment using the Clinical Dementia Rating (CDR) scale. 15 Dementia was diagnosed according to the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD‐10) criteria or a CDR score of 1 to 3. 16 All dementia diagnoses were independently reviewed by a panel of geriatric psychiatrists, with final adjudication by the principal investigator in case of disagreement.

2.4. Assessment of covariates

Demographic characteristics (age, sex, and educational attainment), common ophthalmological conditions (cataract and glaucoma), physical and psychiatric comorbidities (diabetes, hypertension, hypercholesterolemia, heart diseases, obesity, hearing impairment, poor mobility, and depression), and lifestyle factors (physical, intellectual, and social activities, fruit and vegetable intake, and smoking) were examined. Cataract and glaucoma were diagnosed by ophthalmologists. For the present analysis, cataract and glaucoma were examined as binary variables; data on laterality, disease subtype, severity, treatment, and onset were not collected. Other health problems were diagnosed by physicians using ICD‐10 criteria. Obesity was defined as a body mass index of 25 kg/m2 or greater, according to Asian standards. 17 Hearing impairment was defined as hearing loss of ≥40 dB at 1 to 2 kHz in the better ear during audiometric testing (Audioscope, Welch Allyn 23300). Poor mobility was defined as requiring walking aids or being chairbound. Lifestyle behaviors during the preceding month were assessed using a standardized self‐reported questionnaire and categorized according to a locally validated leisure activity classification system. 18 Adequate fruit and vegetable intake and regular participation in physical, intellectual, and social activities were defined as previously reported. 13 , 19 , 20

2.5. Statistical analyses

A sample size of 1000 would have provided ≥80% power (two‐sided α = 0.05) to detect an association between moderate to severe visual impairment and dementia, assuming an annual dementia incidence of 1% and a hazard ratio (HR) of 2.0 or greater. 21

Follow‐up time was calculated in person‐years from baseline until dementia diagnosis, death, loss to follow‐up, or end of follow‐up. Baseline characteristics were compared using independent‐sample t‐tests, one‐way ANOVA, or chi‐squared tests, as appropriate.

Cox proportional hazards models were conducted to examine the associations between baseline visual acuity and incident dementia. Model 1 was unadjusted. Model 2 was adjusted for demographic characteristics, ophthalmological conditions, physical and psychiatric comorbidities, and lifestyle factors. Participants were censored at death, dropout, or end of follow‐up, and HRs with 95% confidence intervals (CIs) were estimated. Given the higher mortality risk associated with poor visual acuity, competing‐risk analyses were additionally performed using Fine–Gray subdistribution hazard models, with death treated as a competing event. Model 3 comprised Model 2 with additional exclusion of participants who developed dementia within the first 6 years of follow‐up or whose dementia‐free status by Year 6 could not be confirmed because of dropout, to minimize potential reverse causation in accordance with the Lancet Commission recommendations for dementia risk‐factor studies. 4 HRs were then estimated for incident dementia occurring during the subsequent 10 years (Years 7 to 16).

Dementia risk was further examined across baseline visual impairment severity, classified using ICD‐11 LogMAR thresholds as normal (<0.3), mild (0.3 to 0.47), moderate (0.48 to 0.99), and severe (≥1.00), applying the same adjusted models and the 6‐year exclusion window. Dementia‐free survival probabilities and corresponding standard errors (SEs) by baseline visual impairment severity were estimated using Kaplan‐Meier methods. The missing data mechanism was assessed using Little's test for missing completely at random. Differences in the timing of dementia onset were quantified using restricted mean survival time (RMST) up to 16 years, calculated as the area under the Kaplan‐Meier survival curve, with RMST differences relative to normal vision used to estimate average differences in dementia‐free years.

To assess whether the observed association remained consistent across age groups and diabetes status, we conducted sensitivity analyses using fully adjusted models stratified by baseline age (65 to 74 years vs ≥75 years) and diabetes status (with vs without diabetes). Continuous age was retained as a covariate in the age‐stratified models, whereas diabetes was omitted from the diabetes‐stratified models because it was constant within each stratum. Effect modification was further assessed by including interaction terms between age group or diabetes status and baseline visual acuity.

To evaluate whether visual acuity level at any given time and its subsequent rate of change were associated with dementia risk, joint models of longitudinal visual acuity and time to dementia onset were applied. These models combined a linear mixed‐effects model for LogMAR visual acuity trajectories (incorporating random intercepts and slopes) and a Cox proportional hazards model for incident dementia, linking the two processes through both the current LogMAR value and its rates of change, including their interaction. The models were adjusted for potential confounders, and only visual acuity measurements obtained prior to dementia onset were included in the analyses. Joint models were fitted using Bayesian estimation, with 10,000 iterations and 2000 burn‐in samples, implemented in R (version 4.4.2) with the JMbayes2 package (version 0.5–0).

3. RESULTS

3.1. Dementia incidence over 16 years

Among the 15,576 participants included in the study, 168,434 person‐years of follow‐up were accrued over 16 years. At baseline, the mean age was 74.5 years (SD = 4.8), and 9945 participants (63.8%) were female (Table 1).

TABLE 1.

Baseline characteristics and 16‐year dementia incidence by visual impairment severity.

Vision
Whole sample (N = 15,576) Normal (n = 1486) Mild impairment (n = 8915) Moderate impairment (n = 4714) Severe impairment (n = 461) p value
Baseline characteristics
Age, mean (SD) 74.5 (4.8) 72.0 (3.8) 73.9 (4.5) 75.9 (5.1) 77.6 (5.8) <0.001
Female, n (%) 9945 (63.8) 736 (49.5) 5662 (63.5) 3218 (68.3) 329 (71.4) <0.001
No formal education received, n (%) 4170 (26.8) 162 (10.9) 2236 (25.1) 1594 (33.8) 178 (38.6) <0.001
Cataract, n (%) 9978 (64.1) 673 (45.3) 5294 (59.4) 3643 (77.3) 368 (79.8) <0.001
Glaucoma, n (%) 544 (3.5) 38 (2.6) 278 (3.1) 190 (4.0) 38 (8.2) <0.001
Diabetes mellitus, n (%) 2358 (15.1) 186 (12.5) 1316 (14.8) 772 (16.4) 84 (18.2) <0.001
Hypertension, n (%) 10,024 (64.4) 922 (62.0) 5658 (63.5) 3131 (66.4) 313 (67.9) <0.001
Hypercholesterolemia, n (%) 6585 (42.3) 640 (43.1) 3800 (42.6) 1958 (41.5) 187 (40.6) 0.49
Obesity, n (%) 5916 (38.0) 537 (36.1) 3395 (38.1) 1810 (38.4) 174 (37.8) 0.46
Heart disease, n (%) 1763 (11.3) 167 (11.2) 975 (10.9) 558 (11.8) 63 (13.7) 0.17
Hearing impairment, n (%) 3532 (22.7) 283 (19.0) 1,904 (21.4) 1209 (25.6) 136 (29.5) <0.001
Poor mobility, n (%) 1156 (7.4) 34 (2.3) 524 (5.9) 497 (10.5) 101 (21.9) <0.001
Depression, n (%) 590 (3.8) 21 (1.4) 333 (3.7) 215 (4.6) 21 (4.6) <0.001
Physical exercises, n (%) 7758 (49.8) 870 (58.5) 4655 (52.2) 2064 (43.8) 169 (36.7) <0.001
Intellectual activities, n (%) 10,202 (65.5) 1216 (81.8) 6010 (67.4) 2785 (59.1) 191 (41.4) <0.001
Social activities, n (%) 12,114 (77.8) 1122 (75.5) 6980 (78.3) 3646 (77.3) 366 (79.4) 0.07
Adequate intake of fruits and vegetables, n (%) 7732 (49.6) 793 (53.4) 4514 (50.6) 2226 (47.2) 199 (43.2) <0.001
Smoking, n (%) 758 (4.9) 60 (4.0) 442 (5.0) 236 (5.0) 20 (4.3) 0.42
Outcome in 16 Years
Incident dementia, n (%) 3643 (23.4) 247 (16.6) 2015 (22.6) 1249 (26.5) 132 (28.6) <0.001
Open in a new tab

A total of 3643 participants (23.4%) developed incident dementia during follow‐up.

3.2. Baseline visual acuity level and risk of dementia

Mean baseline visual acuity for the cohort was 0.46 LogMAR (SD = 0.16), corresponding to approximately 20/58 on the Snellen scale.

Poorer baseline visual acuity level (expressed on the original LogMAR scale) was associated with a higher risk of incident dementia in unadjusted analyses (Model 1, Table 2: unadjusted HR = 3.81; 95% CI = 3.16 to 4.60; p < 0.001). The association remained significant after adjusting for demographic factors, ophthalmological conditions, physical and psychiatric comorbidities, and lifestyle behaviors (Model 2, Table 2: adjusted HR = 1.67; 95% CI = 1.36 to 2.06; p < 0.001). The association persisted, albeit attenuated, after additionally excluding participants who developed dementia within the first 6 years or whose dementia‐free status at Year 6 could not be confirmed (n = 1349) (Model 3, Table 2: adjusted HR = 1.37; 95% CI = 1.04 to 1.80; p = 0.02).

TABLE 2.

Hazard ratios (HRs) and 95% confidence intervals (95% CIs) for incident dementia by baseline visual acuity level and visual impairment severity.

Model 1 Model 2 Model 3
HR (95% CI) p value HR (95% CI) p value HR (95% CI) p value
LogMAR 3.81 (3.16, 4.60) <0.001 1.67 (1.36, 2.06) <0.001 1.37 (1.04, 1.80) 0.02
Normal vision Reference
Mild visual impairment 1.54 (1.35, 1.76) <0.001 1.19 (1.04, 1.36) 0.011 1.20 (0.99, 1.46) 0.06
Moderate visual impairment 2.16 (1.88, 2.47) <0.001 1.37 (1.18, 1.58) <0.001 1.42 (1.16, 1.74) 0.001
Severe visual impairment 2.64 (2.14, 3.26) <0.001 1.41 (1.14, 1.76) 0.002 1.57 (1.12, 2.21) 0.009
Open in a new tab

Model 1: unadjusted.

Model 2: Model 1 + adjusted for demographic factors, ophthalmological conditions, physical and psychiatric comorbidities, and lifestyle behaviors at baseline.

Model 3: Model 2 + excluding dementia cases occurring within the first 6 years of follow–up.

In competing‐risk analyses accounting for death, poorer baseline visual acuity level was still associated with higher risk of dementia (SHR = 1.08; 95% CI = 1.06 to 1.10; p < 0.001).

Neither cataract nor glaucoma was independently associated with higher dementia risk after adjusting for baseline visual acuity, demographic factors, physical and psychiatric comorbidities, and lifestyle behaviors (Model 2, eTable 1) and excluding early dementia cases (Model 3, eTable 1).

3.3. Baseline visual impairment severity and risk of dementia

The proportion of participants who developed incident dementia increased with greater severity of visual impairment at baseline, being highest among those with severe visual impairment (28.6%), followed by moderate (26.5%) and mild visual impairment (22.6%), and lowest among those with normal vision (16.6%) (p < 0.001; Table 1).

In unadjusted analyses, mild, moderate, and severe visual impairments were each associated with higher dementia risk compared with normal vision, with unadjusted HRs of 1.54 (95% CI = 1.35 to 1.76), 2.16 (95% CI = 1.88 to 2.47), and 2.64 (95% CI = 2.14 to 3.26), respectively (all p < 0.001; Model 1, Table 2). However, after adjusting for demographic factors, ophthalmological conditions, physical and psychiatric comorbidities, and lifestyle behaviors and excluding dementia cases occurring within the first 6 years of follow‐up, only moderate to severe visual impairment remained associated with increased dementia risk (Model 3 in Table 2: moderate visual impairment: adjusted HR = 1.42, 95% CI = 1.16 to 1.74, p = 0.001; severe visual impairment: adjusted HR = 1.57, 95% CI = 1.12 to 2.21, p = 0.009).

Dementia‐free survival over 16 years declined progressively with increasing baseline visual impairment severity (Figure 2). Missing data were consistent with being missing completely at random (Little's test, p = 0.36). By Year 16, Kaplan‐Meier‐estimated dementia‐free survival probability decreased from 0.684 (SE = 0.020) among participants with normal vision to 0.594 (SE = 0.009), 0.472 (SE = 0.014), and 0.382 (SE = 0.051) among those with mild, moderate, and severe visual impairment, respectively.

FIGURE 2.

FIGURE 2

Open in a new tab

Dementia‑free survival over 16 years by baseline visual impairment severity. VI, visual impairment.

Consistently, restricted mean dementia‑free survival over 16 years shortened with greater baseline visual impairment severity (14.9, 14.1, 13.4, and 12.8 years, respectively), corresponding to average reductions of 0.8, 1.5, and 2.1 dementia‐free years compared with normal vision.

3.4. Sensitivity analyses by age and diabetes status

In sensitivity analyses, the associations varied by baseline age group (eTable 2). Among participants aged 65 to 74 years, poorer visual acuity was associated with a higher risk of dementia (adjusted HR = 2.19, 95% CI = 1.62 to 2.95), as were moderate (adjusted HR = 1.42, 95% CI = 1.19 to 1.68) and severe visual impairment (adjusted HR 1.49, 95% CI = 1.07 to 2.08). Corresponding estimates were attenuated and no longer statistically significant among participants aged ≥75 years. Formal interaction tests indicated significant effect modification by age for both continuous LogMAR visual acuity (p for interaction < 0.001) and visual impairment severity (p for interaction = 0.008).

Among participants without diabetes, poorer visual acuity, moderate visual impairment, and severe visual impairment were associated with a higher risk of dementia, with adjusted HRs of 1.69 (95% CI = 1.35 to 2.11), 1.35 (95% CI = 1.16 to 1.57), and 1.41 (95% CI = 1.12 to 1.78), respectively. Although these associations were not statistically significant among participants with diabetes, formal interaction tests showed no evidence of effect modification by diabetes status for either LogMAR visual acuity (p for interaction = 0.432) or visual impairment severity (p for interaction = 0.826).

3.5. Visual acuity and subsequent rate of change and risk of dementia

Joint modeling analyses (Figure 3) illustrated how dementia risk varied according to visual acuity level (x‐axis) and its subsequent rate of change (y‐axis).

FIGURE 3.

FIGURE 3

Open in a new tab

Dementia risk prediction by visual acuity level and its subsequent rate of change.

Among participants with stable visual trajectories (slope = 0 LogMAR per year), poorer visual acuity level (expressed per 0.1‐unit increase in LogMAR) was associated with a higher risk of dementia (unadjusted HR = 1.17; 95% CI = 1.13 to 1.21; p < 0.001). The association remained significant after adjusting for confounders (adjusted HR = 1.13; 95% CI = 1.10 to 1.17; p < 0.001). Notably, HRs of 2.0 or greater were observed at visual acuity levels of LogMAR 0.55 or worse.

In contrast, among participants with worsening visual trajectories (slope > 0 LogMAR per year), the association with dementia risk varied by visual acuity level. Worsening trajectories were not significantly associated with a higher risk of dementia until visual acuity levels approached LogMAR 0.61 or worse, beyond which dementia risk increased more sharply in the upper range of the heatmap (Figure 3).

Associations for improving visual trajectories were also examined. Compared with a worsening visual trajectory (slope = 0.02 LogMAR per year; 97.5th percentile), which was associated with a 24% increase in dementia risk per 0.1‐unit increase in LogMAR, an improving visual trajectory (slope = −0.01 LogMAR per year; 2.5th percentile) was associated with a more modest 14% increase in dementia risk per 0.1‐unit increase in LogMAR.

4. DISCUSSION

In this 16‐year, territory‐wide cohort study of more than 15,000 community‐dwelling older adults, poorer visual acuity level was associated with a higher risk of incident dementia, independently of demographic, health, and lifestyle factors and after exclusion of early dementia cases. Notably, moderate to severe, but not mild, visual impairment was independently associated with increased dementia risk. Participants with moderate and severe visual impairment developed dementia approximately 1.5 and 2.1 years earlier, respectively, than those with normal vision. Moreover, among these individuals, their risk of dementia was further amplified in the presence of a rapid decline in visual acuity. Improvements in visual acuity were associated with a dementia risk that was not as high as that seen with continued visual deterioration, although the risk was not fully eliminated. Taken together, these findings strengthen the evidence supporting visual impairment as a dementia risk factor and highlight the importance of incorporating visual assessment with longitudinal monitoring into dementia risk stratification.

4.1. Comparison with previous studies

Compared with prior longitudinal studies, 22 , 23 , 24 , 25 , 26 , 27 our study benefits from extended follow‐up, rigorous clinical ascertainment of dementia, and comprehensive control of confounding factors, thereby reducing the likelihood of reverse directionality, misclassification error, and confounding effects. To our knowledge, this is among the largest and longest community‐based cohort studies to demonstrate an independent association between visual acuity and dementia risk. The long follow‐up period strengthens the temporal link between visual impairment and dementia onset. Although causality cannot be inferred from this observational design, the observation that visual impairment precedes dementia onset by a significant margin supports its role as a risk factor rather than solely a prodromal manifestation. Together with the graded associations across severity levels observed in the present study, where moderate to severe, but not mild, visual impairment is independently associated with higher dementia incidence, these results provide additional evidence for a possible dose‐dependent relationship between visual impairment and dementia risk.

Our results also highlight a more complex relationship between visual impairment and dementia. In contrast to our earlier 6‐year analysis, in which poorer baseline visual acuity, but not the rate of decline, was associated with increased dementia risk, 21 the extended follow‐up revealed associations with both acuity level and trajectory. This pattern is unlikely to be attributable to the large sample size alone, as similar associations with trajectories were not observed at 6 years. Instead, the emergence of these associations over extended follow‐up likely reflects a greater number of dementia events occurred and longer exposure to divergence in visual decline. Importantly, two clinically relevant thresholds were identified: Dementia risk approximately doubled at a visual acuity of LogMAR 0.55 even among individuals with stable vision, and the impact of visual decline on dementia risk became more pronounced at LogMAR levels of approximately 0.61 or worse. These findings suggest that moderate to severe visual impairment confers higher dementia risk, which may be further exacerbated by rapid visual decline. Our results highlight the importance of healthcare professionals recognizing and addressing increased dementia risk in older adults with impaired and worsening vision. Regular eye assessments may therefore be warranted, as monitoring both visual acuity and its longitudinal changes could help identify individuals at increased risk.

Contrary to some previous reports, 28 , 29 we did not observe independent associations between cataract or glaucoma and dementia risk after accounting for visual acuity. However, these findings should be interpreted with caution. Although ophthalmologist‐confirmed diagnoses were available, information was limited to the presence or absence of each condition, without details on laterality, subtype, severity, duration, or treatment. As these characteristics may differentially affect visual function and dementia risk, we were unable to assess potential heterogeneity across disease subgroups. Consequently, associations specific to particular forms or stages of cataract or glaucoma may have been attenuated and therefore not detected in the present analyses. Prior studies reported improvements in cognitive outcomes and structural brain changes following cataract surgery. 30 , 31 , 32 , 33 , 34 In this context, our observation that improving visual acuity was associated with attenuated dementia risk is consistent with the hypothesis that addressing visual impairment may moderate, though not fully reverse, dementia risk. Long‐term intervention studies will be required to determine whether correcting visual impairment, even in late life, can delay or prevent the clinical manifestation of dementia.

Realizing any potential benefit from earlier identification and treatment of visual impairment will depend on access to and uptake of appropriate eye care, including regular eye examinations, timely spectacle correction, and low‐vision rehabilitation. Studies among older adults have identified low perceived need, limited understanding of preventive eye care, and practical and service‐related barriers to preventive service utilization, including routine eye examinations. 35 , 36 These findings suggest that screening alone may be insufficient without accessible services, effective communication, and support for follow‐up care. A comprehensive approach spanning early detection, access to eye care, and ongoing management of visual impairment may be required to maximize any potential benefits for brain health in later life.

Although the mechanisms underlying these associations remain to be elucidated, our findings provide several insights. The persistence of the association after adjustment for age suggests that visual impairment is not merely a surrogate for chronological aging. At the same time, the significant interaction by age indicates that the association was stronger among younger‐old adults (65 to 74 years) than among older‐old adults (≥75 years). The attenuated association in the latter group may reflect survivor effects, greater heterogeneity, reduced statistical power, or the increasing influence of competing age‐related disease burden in advanced age. From a clinical perspective, these findings highlight the potential importance of timely identification and management of visual impairment in later life, particularly among younger‐old adults.

Although the associations were more evident among participants without diabetes, formal interaction testing provided no evidence that the association differed by diabetes status. This finding suggests that the observed relationship is not solely attributable to diabetes‐related vascular or metabolic disease burden and may be relevant across different metabolic risk profiles. Nevertheless, diabetes was classified only according to its presence or absence at baseline in the present analysis. Future studies incorporating information on diabetes onset, duration, complications, glycemic control, and treatment may help further clarify these relationships.

Several biological and behavioral mechanisms may underlie the association between visual impairment and dementia development. Diminished visual input and sensory deprivation may contribute to downstream neurodegeneration through reduced cortical stimulation. Consistent with this hypothesis, neuroimaging studies have demonstrated associations between visual impairment and structural and functional changes beyond the visual cortex, including the hippocampus and frontal networks. 37 , 38 , 39 , 40 , 41 , 42 Visual impairment may also influence dementia risk indirectly through increased falls, reduced mobility and physical activity, social isolation, loneliness, depression, and reduced participation in cognitively stimulating activities, many of which are established risk factors for dementia. Although some of these factors were adjusted for at baseline, such adjustment does not preclude the possibility that visual impairment contributes to dementia risk through subsequent changes in mobility, activity levels, social engagement, or psychological well‐being over time.

4.2. Strengths and limitations

This study has several strengths. It followed a large, territory‐wide community cohort for over a decade, with a moderate but acceptable attrition rate given the age of participants and the extended follow‐up. The clinical diagnosis of dementia was established through comprehensive and detailed clinical assessments conducted independently of visual function, minimizing misclassification. Individuals who developed dementia within the first 6 years were excluded from the analysis, minimizing the possibility of reverse directionality. Additionally, this study accounted for a wide range of potential confounders, including physical health, sociodemographic factors, and health‐related behaviors.

Nevertheless, this study has some limitations. First, the observational design precludes causal inference. Although the extended temporal separation between visual impairment and dementia onset, the observed dose–response relationship, and the persistence of associations after exclusion of early dementia cases reduce the likelihood of reverse causation, residual confounding from unmeasured factors cannot be excluded. Visual impairment may partly reflect accelerated biological aging, frailty, vascular burden, or shared neurodegenerative processes that increase the likelihood of both visual decline and dementia. Second, despite active tracing and registry linkage, some under‐ascertainment of dementia among participants lost to follow‐up may have occurred. Third, habitual, rather than best‐corrected, high‐contrast distance visual acuity was assessed. While this approach may better reflect real‐world risk exposure, it does not fully distinguish visual impairment attributable to uncorrected or undercorrected refractive error from that caused by ocular disease and captures only one dimension of visual function. Future studies incorporating a broader range of visual function measures, including near visual acuity, contrast sensitivity, and visual fields, may provide additional insights into the relationship between visual impairment and dementia risk. Fourth, data on earlier‐life cognition and detailed information on the onset, severity, and treatment of cataract and glaucoma were unavailable. Other age‐related eye diseases, including age‐related macular degeneration and diabetic retinopathy, were also not examined. These conditions may affect visual function through different pathophysiological mechanisms and may have distinct associations with cognitive outcomes. Fifth, genetic factors (e.g., ApoE4) and blood‐based or neuroimaging biomarkers were not available, precluding subtype‐specific analyses. Sixth, fewer than 10% of participants had normal vision at baseline, and fewer than 17% showed improvement in visual acuity, limiting precision at the extremes of visual change. Finally, generalizability to other populations should be approached with caution, as the cohort comprised exclusively older, community‐dwelling ethnic Chinese adults who had a relatively low burden of comorbidities.

4.3. Conclusions

In this large, long‐term cohort study, poorer visual acuity was independently associated with an increased risk of dementia. Moderate to severe visual impairment was linked to earlier dementia onset, with risk further amplified by rapid subsequent visual decline. The substantial temporal separation between visual impairment and dementia onset strengthens the case for visual impairment as a risk factor rather than solely a prodromal symptom. These findings underscore the importance of incorporating visual assessments into dementia risk stratification and highlight the need for future studies to determine whether timely identification and correction of visual impairment can contribute to dementia prevention.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest. Author disclosures are available in the Supporting Information.

CONSENT STATEMENT

Written informed consent was obtained from all participants or from their family members or legal guardians if they became mentally incapable of giving consent before the follow‐up assessment. Ethics approval was granted by the Ethics Committees of the Department of Health of the Government of the Hong Kong SAR and the Chinese University of Hong Kong. This study was performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.

Supporting information

Supporting Information

DAD2-18-e70499-s001.docx (19.1KB, docx)

Supporting Information

DAD2-18-e70499-s002.pdf (371KB, pdf)

ACKNOWLEDGMENTS

We thank the staff of the 18 EHCs and all participants and their family members for their time and contribution to this study. We also thank the Elderly Health Service from the Department of Health of the Government of the Hong Kong SAR for collaboration and assistance with mortality record linkage through the Death Registry. We are grateful to the research assistants of the Dementia Research Unit from the Department of Psychiatry at the Chinese University of Hong Kong for their support in tracing participants lost to follow‐up. This study was supported by the Health and Medical Research Fund (HMRF) of the Health Bureau of the Government of the Hong Kong SAR (grant number: CFS‐CUHK3). The funder had no role in the study design; data collection, analysis, or interpretation; manuscript preparation; or the decision to submit the article for publication.

Lee ATC, Leung ONW, Huo Z, et al. Severity and progression of visual impairment and risk of dementia. Alzheimer's Dement. 2026;18:e70499. 10.1002/dad2.70499

Allen T. C. Lee and Owen N. W. Leung are Co‐first authors.

REFERENCES

  • 1. Varma R, Vajaranant TS, Burkemper B, et al. Visual impairment and blindness in adults in the United States: demographic and geographic variations from 2015 to 2050. JAMA Ophthalmol. 2016;134(7):802‐809. doi: 10.1001/jamaophthalmol.2016.1284 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Chan T, Friedman DS, Bradley C, Massof R. Estimates of incidence and prevalence of visual impairment, low vision, and blindness in the United States. JAMA Ophthalmol. 2018;136(1):12‐19. doi: 10.1001/jamaophthalmol.2017.4655 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3. Wagner LD, Rein DB. Attributes associated with eye care use in the United States: a meta‐analysis. Ophthalmology. 2013;120(7):1497‐1501. doi: 10.1016/j.ophtha.2012.12.030 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. 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. doi: 10.1016/s0140-6736(24)01296-0 [DOI] [PubMed] [Google Scholar]
  • 5. Vu TA, Fenwick EK, Gan ATL, et al. The bidirectional relationship between vision and cognition: a systematic review and meta‐analysis. Ophthalmology. 2021;128(7):981‐992. doi: 10.1016/j.ophtha.2020.12.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6. 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 [DOI] [PubMed] [Google Scholar]
  • 7. Kuźma E, Littlejohns TJ, Khawaja AP, Llewellyn DJ, Ukoumunne OC, Thiem U. Visual impairment, eye diseases, and dementia risk: a systematic review and meta‐analysis. J Alzheimers Dis. 2021;83(3):1073‐1087. doi: 10.3233/jad-210250 [DOI] [PubMed] [Google Scholar]
  • 8. Nagarajan N, Assi L, Varadaraj V, et al. Vision impairment and cognitive decline among older adults: a systematic review. BMJ Open. 2022;12(1):e047929. doi: 10.1136/bmjopen-2020-047929 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Nelson PA, Dial JG, Joyce A. Validation of the cognitive test for the blind as an assessment of intellectual functioning. Rehabilitation Psychology. 2002;47(2):184‐193. doi: 10.1037/0090-5550.47.2.184 [DOI] [Google Scholar]
  • 10. Dawes P, Reeves D, Yeung WK, et al. Development and validation of the Montreal Cognitive Assessment for people with vision impairment (MOCA‐VI). Psychol Assess. 2025;37(3):114‐122. doi: 10.1037/pas0001357 [DOI] [PubMed] [Google Scholar]
  • 11. Liao H, Zhu Z, Peng Y. Potential utility of retinal imaging for Alzheimer's disease: a review. Front Aging Neurosci. 2018;10:188. doi: 10.3389/fnagi.2018.00188 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Chiquita S, Rodrigues‐Neves AC, Baptista FI, et al. The retina as a window or mirror of the brain changes detected in Alzheimer's disease: critical aspects to unravel. Mol Neurobiol. 2019;56(8):5416‐5435. doi: 10.1007/s12035-018-1461-6 [DOI] [PubMed] [Google Scholar]
  • 13. Lee ATC, Richards M, Chan WC, Chiu HFK, Lee RSY, Lam LCW. Association of daily intellectual activities with lower risk of incident dementia among older Chinese adults. JAMA Psychiatry. 2018;75(7):697‐703. doi: 10.1001/jamapsychiatry.2018.0657 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. Chu LW, Pei CKW, Ho MH, Chan PT. Validation of the Abbreviated Mental Test (Hong Kong version) in the elderly medical patient. Hong Kong Medical Journal. 1995;1(3):207‐211. https://www.hkmj.org/system/files/hkm9509p207.pdf [Google Scholar]
  • 15. Morris JC. Clinical dementia rating: a reliable and valid diagnostic and staging measure for dementia of the Alzheimer type. Int Psychogeriatr. 1997;9(Suppl 1):173‐176. doi: 10.1017/s1041610297004870 [DOI] [PubMed] [Google Scholar]
  • 16. Lim WS, Chin JJ, Lam CK, Lim PP, Sahadevan S. Clinical Dementia Rating: experience of a multi‐racial Asian population. Alzheimer Dis Assoc Disord. 2005;19(3):135‐142. doi: 10.1097/01.wad.0000174991.60709.36 [DOI] [PubMed] [Google Scholar]
  • 17. Expert Consultation WHO. Appropriate body‐mass index for Asian populations and its implications for policy and intervention strategies. Lancet. 2004 10;363(9403):157‐163. doi: 10.1016/s0140-6736(03)15268-3 [DOI] [PubMed] [Google Scholar]
  • 18. Leung GTY, Leung KF, Lam LCW. Classification of late‐life leisure activities among elderly Chinese in Hong Kong. East Asian Arch Psychiatry. 2011;21(3):123‐127. https://www.easap.asia/index.php/component/k2/item/180‐v21n3‐p123‐127 [PubMed] [Google Scholar]
  • 19. Lee ATC, Richards M, Chan WC, Chiu HFK, Lee RSY, Lam LCW. Lower risk of incident dementia among Chinese older adults having three servings of vegetables and two servings of fruits a day. Age Ageing. 2017;46(5):773‐779. doi:10.1093/ageing/afx018 [DOI] [PubMed] [Google Scholar]
  • 20. Lee ATC, Richards M, Chan WC, Chiu HFK, Lee RSY, Lam LCW. Intensity and types of physical exercise in relation to dementia risk reduction in community‐living older adults. J Am Med Dir Assoc. 2015;16(10):899.e1‐899.e7. doi:10.1016/j.jamda.2015.07.012 [DOI] [PubMed] [Google Scholar]
  • 21. Lee ATC, Richards M, Chan WC, Chiu HFK, Lee RSY, Lam LCW. Higher dementia incidence in older adults with poor visual acuity. J Gerontol A Biol Sci Med Sci. 2020;75(11):2162‐2168. doi:10.1093/gerona/glaa036 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Davies‐Kershaw HR, Hackett RA, Cadar D, Herbert A, Orrell M, Steptoe A. Vision impairment and risk of dementia: findings from the English Longitudinal Study of Ageing. J Am Geriatr Soc. 2018;66(9):1823‐1829. doi:10.1111/jgs.15456 [DOI] [PubMed] [Google Scholar]
  • 23. Brenowitz WD, Kaup AR, Lin FR, Yaffe K. Multiple sensory impairment is associated with increased risk of dementia among black and white older adults. J Gerontol A Biol Sci Med Sci. 2019;74(6):890‐896. doi:10.1093/gerona/gly264 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24. Rogers MA, Langa KM. Untreated poor vision: a contributing factor to late‐life dementia. Am J Epidemiol. 2010;171(6):728‐735. doi:10.1093/aje/kwp453 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25. Naël V, Pérès K, Dartigues JF, et al. Vision loss and 12‐year risk of dementia in older adults: the 3C cohort study. Eur J Epidemiol. 2019;34(2):141‐152. doi:10.1007/s10654‐018‐00478‐y [DOI] [PubMed] [Google Scholar]
  • 26. Ehrlich JR, Swenor BK, Zhou Y, Langa KM. The longitudinal association of vision impairment with transitions to cognitive impairment and dementia: findings from the Aging, Demographics and Memory Study. J Gerontol A Biol Sci Med Sci. 2021;76(12):2187‐2193. doi:10.1093/gerona/glab157 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27. Littlejohns TJ, Hayat S, Luben R, et al. Visual impairment and risk of dementia in 2 population‐based prospective cohorts: UK Biobank and EPIC‐Norfolk. J Gerontol A Biol Sci Med Sci. 2022;77(4):697‐704. doi:10.1093/gerona/glab325 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28. Wen Y‐H, Wu S‐S, Lin C‐HR, et al. A Bayesian approach to identifying new risk factors for dementia: a nationwide population‐based study. Medicine (Baltimore). 2016;95(21):e3658. doi:10.1097/md.0000000000003658 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Lee CS, Larson EB, Gibbons LE, et al. Associations between recent and established ophthalmic conditions and risk of Alzheimer's disease. Alzheimers Dement. 2019;15(1):34‐41. doi:10.1016/j.jalz.2018.06.2856 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Sanda N, Cerliani L, Authié CN, et al. Visual brain plasticity induced by central and peripheral visual field loss. Brain Struct Funct. 2018;223(7):3473‐3485. doi:10.1007/s00429‐018‐1700‐7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. Lou AR, Madsen KH, Julian HO, et al. Postoperative increase in grey matter volume in visual cortex after unilateral cataract surgery. Acta Ophthalmol. 2013;91(1):58‐65. doi:10.1111/j.1755‐3768.2011.02304.x [DOI] [PubMed] [Google Scholar]
  • 32. Fukuoka H, Sutu C, Afshari NA. The impact of cataract surgery on cognitive function in an aging population. Curr Opin Ophthalmol. 2016;27(1):3‐8. doi:10.1097/icu.0000000000000226 [DOI] [PubMed] [Google Scholar]
  • 33. Lin H, Zhang Li, Lin D, et al. Visual restoration after cataract surgery promotes functional and structural brain recovery. EBioMedicine. 2018;30:52‐61. doi:10.1016/j.ebiom.2018.03.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. Maharani A, Dawes P, Nazroo J, Tampubolon G, Pendleton N. SENSE‐Cog WP1 group. Cataract surgery and age‐related cognitive decline: a 13‐year follow‐up of the English Longitudinal Study of Ageing. PLoS One. 2018;13(10):e0204833. doi:10.1371/journal.pone.0204833 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Lau WY, Lian J, Lam A, Yap M. A theory‐ and evidence‐based holistic intervention to enhance the uptake of preventive eye examination among the elderly centre members. Public Health Pract (Oxf). 2025;10:100660. doi:10.1016/j.puhip.2025.100660 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. Liao Q, Lau W, McGhee S, et al. Barriers to preventive care utilization among Hong Kong community‐dwelling older people and their views on using financial incentives to improve preventive care utilization. Health Expect. 2021;24(4):1242‐1253. doi:10.1111/hex.13256 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. Leporé N, Shi Y, Lepore F, et al. Pattern of hippocampal shape and volume differences in blind subjects. Neuroimage. 2009;46(4):949‐957. doi:10.1016/j.neuroimage.2009.01.071 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38. Chen WW, Wang N, Cai S, et al. Structural brain abnormalities in patients with primary open‐angle glaucoma: a study with 3T MR imaging. Invest Ophthalmol Vis Sci. 2013;54(1):545‐554. doi:10.1167/iovs.12‐9893 [DOI] [PubMed] [Google Scholar]
  • 39. Frezzotti P, Giorgio A, Motolese I, et al. Structural and functional brain changes beyond visual system in patients with advanced glaucoma. PLoS One. 2014;9(8):e105931. doi:10.1371/journal.pone.0105931 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Wang J, Li T, Sabel BA, et al. Structural brain alterations in primary open angle glaucoma: a 3T MRI study. Sci Rep. 2016;6:18969. doi:10.1038/srep18969 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41. Sabbah N, Sanda N, Authié CN, et al. Reorganization of early visual cortex functional connectivity following selective peripheral and central visual loss. Sci Rep. 2017;7:43223. doi:10.1038/srep43223 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42. Nuzzi R, Dallorto L, Rolle T. Changes of visual pathway and brain connectivity in glaucoma: a systematic review. Front Neurosci. 2018;12:363. doi:10.3389/fnins.2018.00363 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting Information

DAD2-18-e70499-s001.docx (19.1KB, docx)

Supporting Information

DAD2-18-e70499-s002.pdf (371KB, pdf)

Articles from Alzheimer's & Dementia : Diagnosis, Assessment & Disease Monitoring are provided here courtesy of Wiley

RESOURCES