Key Points
Question
Should vision impairment be included in a life-course model of potentially modifiable dementia risk factors?
Findings
In this cross-sectional study of 2018 data from the Health and Retirement Study on 16 690 US adults aged 50 years and older, an estimated 1.8% of dementia cases in the US (more than 100 000 prevalent cases) were associated with vision impairment.
Meaning
These results suggest that vision impairment should be included life-course models of potentially modifiable dementia risk factors.
Abstract
Importance
Dementia prevention is a high priority, given the large impact of dementia on the well-being of individuals and society. The number of older adults with dementia in the US and globally is projected to increase as a result of population aging and growth. Thus, it is vital to identify potentially modifiable dementia risk factors. Vision impairment has been identified as a risk factor for accelerated cognitive decline and incident dementia. An estimated 90% of vision impairment is preventable or has yet to be treated. Nevertheless, vision impairment has not been included in the dominant life-course models of dementia risk factors used to shape public health policy and research priorities.
Objective
To strengthen an existing model of potentially modifiable dementia risk factors through the inclusion of vision impairment and to estimate the contributions of those risk factors in the US population.
Design, Setting, and Participants
Population-based, cross-sectional study using data from the 2018 round of the Health and Retirement Study. Analyses were conducted from March 11 through September 24, 2021. The study population was a probability sample of US adults aged 50 years and older.
Exposures
Potentially modifiable dementia risk factors, including vision impairment.
Main Outcomes and Measures
The estimated population attributable fractions (PAFs) of dementia associated with vision impairment and other dementia risk factors were calculated. The PAF represents the number of cases of dementia that would potentially be prevented if a risk factor were eliminated.
Results
The probability sample from the Health and Retirement Study included 16 690 participants (weighted demographic characteristics: 54.0% female, 52.0% age ≥65, 10.6% Black, 80% White, and 9.2% identified as other [including American Indian or Alaska Native, Asian, and Hawiian Native or Pacific Islander, although specific data were not available]). The 12 dementia risk factors in the PAF model were associated with an estimated 62.4% of dementia cases in the US. The risk factor with the highest weighted PAF for dementia was hypertension (12.4%). The PAF of vision impairment was 1.8%, suggesting that more than 100 000 prevalent dementia cases in the US could potentially have been prevented through healthy vision.
Conclusions and Relevance
Existing life-course models of potentially modifiable dementia risk factors may consider including vision impairment. Since a large majority of vision impairment can be treated with cost-effective but underused interventions, this may represent a viable target for future interventional research that aims to slow cognitive decline and prevent incident dementia.
Using data from the Health and Retirement Study, this cross-sectional study assesses whether vision impairment should be included in a life-course model of potentially modifiable dementia risk factors.
Introduction
The number of US individuals aged 65 years and older with dementia is projected to more than double by 2050, to 13.8 million from 5.8 million in 2020.1,2 The annual economic cost of dementia care in the US is estimated to be more than $500 billion.2,3 Thus, there is an urgent need to identify modifiable risk factors for dementia that can be targeted with interventions to slow cognitive decline and prevent dementia.
The 2020 Lancet Commission report on dementia prevention, intervention, and care proposed a life-course model of 12 potentially modifiable dementia risk factors, including less education, hearing loss, traumatic brain injury, hypertension, excessive alcohol consumption, obesity, smoking, depression, social isolation, physical inactivity, diabetes, and air pollution.4 Together, these factors are associated with about 40% of dementia cases globally.4 Vision impairment was not included in this model, despite considerable evidence that it is associated with an elevated risk of incident dementia5,6,7 and that it may operate through the same pathways as hearing loss,8 the risk factor with the highest population attributable fraction (PAF; the proportion of dementia cases that would potentially be prevented if a risk factor was eliminated) in the Lancet model. These hypothesized pathways include common etiologic factors (eg, neurodegenerative or vascular conditions), increased cognitive load, reduced cognitive stimulation, direct alteration of brain structure, depression, social isolation, and/or physical inactivity.8 Existing evidence suggests that vision impairment exerts its effects primarily in later life (65 years of age or older),9 in contrast to some other risk factors that appear to exert their effect primarily in early life or midlife.
Of note, about 80% of vision impairment occurs in adults aged 50 years and older, and 90% of cases are preventable or have yet to be treated.10 Most cases are treatable with 2 highly cost-effective interventions, eyeglasses and/or cataract surgery. Given that vision impairment is often modifiable10 and, in some cases, related to neurodegenerative processes,11 we sought to determine the PAFs for the potentially modifiable dementia risk factors in the US after the addition of vision impairment to the existing life-course model.
Methods
This cross-sectional study extends the analysis of the PAF of dementia risk factors conducted for the 2020 Lancet Commission report.4 We followed similar methods but used the Health and Retirement Study (HRS) to calculate the prevalence and communality of risk factors; estimates are therefore most relevant to the US population. As an analysis of publicly available, deidentified data, it was exempt from institutional review board approval. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline. Data analyses were performed from March 11 to September 24, 2021.
Study Sample
The HRS is a panel study that surveys more than 20 000 adults 50 years of age and older every 2 years.12 The current analysis used data from the 2018 HRS, which included 16 690 respondents.13 Sample weights were used to estimate the national prevalence of risk factors and to calculate communalities. Since visual difficulty is self-reported in the HRS, we used prevalence data from a recent bayesian meta-analysis on the prevalence of vision impairment in US adults.14 Variable definitions are provided in the eMethods in the Supplement. Although air pollution is included in the Lancet model, we did not include it in our model because those data were not readily available in the HRS.
Statistical Analysis
The PAF is based on the population prevalence and relative risk of dementia for each risk factor. It is weighted based on a principal components analysis (eTable 1 in the Supplement) to account for communality (clustering of risk factors). The detailed methods for calculating the PAF are described in the eMethods in the Supplement. We accounted for missing data through full information maximum likelihood estimation. We also performed sensitivity analyses with effect sizes and prevalence estimates for vision impairment as detailed in the eMethods in the Supplement.
Results
The probability sample included 16 690 participants (weighted demographic characteristics: 54.0% female, 46% male, 51.5% age ≥65, 10.6% Black, 80.2% White, and 9.2% other; although the category “other” specifies American Indian or Alaska Native, Asian, and Native Hawaiian or Pacific Islander, the publicly available files for each wave of the HRS include data only for the categories Black, White, and other) (Table 1). Overall, the 12 potentially modifiable risk factors in our model were associated with an estimated 62.4% of dementia cases in the US. Table 2 shows the weighted and unweighted PAFs for dementia of each risk factor. Hypertension was the most prevalent risk factor (59.6%) and had the highest weighted PAF (12.4). Vision impairment, with a prevalence of 8.3%, had a weighted PAF of 1.8%, suggesting that more than 100 000 prevalent dementia cases in the US could potentially have been prevented through healthy vision. The weighted and unweighted PAFs for each risk factor are shown in the eFigure in the Supplement. In analyses stratified by race and ethnicity, the weighted PAF for vision impairment ranged from 1.8% among White non-Hispanic older adults to 2.2% among Hispanic older adults (eTable 2 in the Supplement). In a sensitivity analysis using different estimates of prevalence and effect size, the overall PAF for dementia of vision impairment ranged from 1.4% to 8.5% (eTable 3 in the Supplement).
Table 1. Demographic and Clinical Characteristics of the Study Population.
| Characteristic | Participants, No. (weighted %) | Missing participants, No. (%) |
|---|---|---|
| Sex | ||
| Female | 9763 (54.0) | NA |
| Male | 6926 (46.0) | 1 (<0.1) |
| Age, y | ||
| 50-64 | 7702 (48.5) | NA |
| ≥65 | 8988 (51.5) | NA |
| Race | ||
| Black | 3660 (10.6) | NA |
| White | 11011 (80.2) | 139 (0.8) |
| Othera | 1880 (9.2) | NA |
| Risk factors | ||
| Poor vision | 1960 (21.4) | 6 (<0.1) |
| Less education | 2770 (11.6) | 926 (5.5) |
| Hearing loss | 3544 (19.2) | 24 (0.1) |
| TBIb | 199b (18.7) | 3 (0.3) |
| Hypertension | 10 445 (59.6) | 20 (0.1) |
| Excessive alcohol consumption | 390 (2.8) | 27 (0.2) |
| Obesity | 6072 (40.3) | 1917 (11.5) |
| Smoking | 2147 (12.0) | 82 (0.5) |
| Depression | 4492 (26.7) | 52 (0.3) |
| Social isolationb | 450b (7.1) | NA |
| Physical inactivity | 3498 (17.4) | 77 (0.5) |
| Diabetes | 4691 (24.4) | 21 (0.1) |
Abbreviations: HRS, Health and Retirement Study; NA, not applicable; TBI, traumatic brain injury.
Although the category “other” in the HRS specifies American Indian or Alaska Native, Asian, and Native Hawaiian or Pacific Islander, the publicly available files for each wave of the study include data only for the categories Black, White, and other.
Not assessed in the entire HRS sample; TBI was determined from a 2014 HRS module focused on this risk factor, and social isolation was determined from a 2018 leave-behind survey from the HRS.
Table 2. Potentially Modifiable Dementia Risk Factors.
| Risk factor | Dementia, RR (95% CI) | Risk factor prevalence, % | Communality, % | PAF, % | |
|---|---|---|---|---|---|
| Unweighted | Weighted | ||||
| Vision impairment | 1.5 (1.4-1.6) | 8.3 | 45.3 | 3.9 | 1.8 |
| Less education | 1.6 (1.3-2.0) | 11.6 | 34.8 | 6.5 | 3.1 |
| Hearing loss | 1.9 (1.4-2.7) | 19.2 | 26.4 | 14.7 | 7.0 |
| TBI | 1.8 (1.5-2.2) | 18.7 | 3.9 | 13.0 | 6.1 |
| Hypertension | 1.6 (1.2-2.2) | 59.6 | 47.8 | 26.3 | 12.4 |
| Excessive alcohol consumption | 1.2 (1.1-1.3) | 2.8 | 3.7 | 0.56 | 0.3 |
| Obesity | 1.6 (1.3-1.9) | 40.3 | 26.0 | 19.5 | 9.2 |
| Smoking | 1.6 (1.2-2.2) | 12.0 | 15.6 | 6.7 | 3.2 |
| Depression | 1.9 (1.6-2.3) | 26.7 | 19.1 | 19.4 | 9.1 |
| Social isolation | 1.6 (1.3-1.9) | 7.1 | 15.5 | 4.1 | 1.9 |
| Physical inactivity | 1.4 (1.2-1.7) | 17.4 | 41.9 | 6.5 | 3.1 |
| Diabetes | 1.5 (1.3-1.8) | 24.4 | 50.3 | 10.9 | 5.1 |
| Overall | NA | NA | NA | NA | 62.4 |
Abbreviations: NA, not applicable; PAF, population attributable fraction; RR, relative risk; TBI, traumatic brain injury.
Discussion
In this cross-sectional study of data from the HRS, we investigated the contribution of vision impairment to a life-course model of potentially modifiable dementia risk factors in the US. We found that vision impairment was associated with a substantial fraction of dementia cases. Specifically, we estimated that about 1.8%, or more than 100 000 cases, of prevalent dementia in the US were associated with vision impairment and estimate that, given current projections for the coming decades,1,2 this number will increase to approximately 250 000 by 2050. Since about 9 in 10 cases of vision impairment are preventable or have yet to be treated with proven and cost-effective interventions,10 vision impairment may represent an important modifiable risk factor.
Of note, the role of vision impairment as a dementia risk factor is even greater when there is a high prevalence of vision impairment. In this study, the PAF was highest (2.2%) among Hispanic older adults in whom the prevalence of vision impairment was 11.0% (compared to 8.3% overall). Likewise, in China and India, the estimated prevalence of vision impairment is 3-fold and 6-fold greater, respectively, than in the US.15 Addressing preventable and treatable causes of vision impairment may therefore represent a novel avenue to slow cognitive decline and prevent dementia for a large number of older adults both in the US and globally. Although vision impairment was not among the risk factors in our model with the highest PAF, the modifiable nature of most vision impairment, coupled with the high clinical and cost-effectiveness of interventions to optimize vision, may make this a particularly attractive intervention target.
While elimination of vision impairment may not alter the prevalence of amyloidopathy or tauopathy associated with dementia, vision impairment may play a role in the onset and progression of symptoms associated with the clinical syndrome of dementia. In this case, prevention or treatment of vision impairment could reduce the prevalence of dementia, even without affecting biomarker-associated diseases that cause dementia. Thus, additional research is needed to test the effectiveness of interventions to preserve cognitive health by promoting healthy vision. Notably, vision improving interventions, including eyeglasses and cataract surgery, remain underused in the US and globally, particularly in disadvantaged communities.10
The relative importance of some risk factors in our model diverged from the Lancet Commission findings4 because of our use of data from the US population. Findings from individual countries may be most relevant to planning contextually relevant public health and policy programs. Future cross-national comparisons could shed light on the relative importance of risk factors across settings.
Limitations
There were several limitations to this study. First, the meta-analyses that estimated effect sizes for dementia risk factors (including vision impairment) did not consistently adjust for all other dementia risk factors.14 Second, we were unable to assess whether the elimination of risk factors may be associated with dementia incidence.
Conclusions
This cross-sectional study estimated that more than 100 000 dementia cases in the US were associated with vision impairment. Since a large majority of vision impairment is modifiable, further investigations are warranted to determine whether interventions to optimize vision will slow cognitive decline and prevent future cases of dementia.
eMethods.
eTable 1. Principal-Component Analysis of Modifiable Risk Factors for Dementia in the Health and Retirement Study
eTable 2. Potentially Modifiable Dementia Risk Factors by Race and Ethnicity
eTable 3. Sensitivity Analyses for Population Attributable Fraction of Dementia Due to Vision Impairment
eFigure. Population Attributable Fraction of Potentially Modifiable Dementia Risk Factors
eReferences
References
- 1.Plassman BL, Langa KM, Fisher GG, et al. Prevalence of dementia in the United States: the Aging, Demographics, and Memory Study. Neuroepidemiology. 2007;29(1-2):125-132. doi: 10.1159/000109998 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Alzheimer’s Association . 2020 Alzheimer’s disease facts and figures. Alzheimers Dement. 2020;15(3):321-387. [Google Scholar]
- 3.Jutkowitz E, Kane RL, Gaugler JE, MacLehose RF, Dowd B, Kuntz KM. Societal and family lifetime cost of dementia: implications for policy. J Am Geriatr Soc. 2017;65(10):2169-2175. doi: 10.1111/jgs.15043 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Livingston G, Huntley J, Sommerlad A, et al. Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet. 2020;396(10 248):413-446. doi: 10.1016/S0140-6736(20)30367-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.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]
- 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.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] [PubMed] [Google Scholar]
- 8.Whitson HE, Cronin-Golomb A, Cruickshanks KJ, et al. American Geriatrics Society and National Institute on Aging Bench-to-Bedside Conference: sensory impairment and cognitive decline in older adults. J Am Geriatr Soc. 2018;66(11):2052-2058. doi: 10.1111/jgs.15506 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Paik JS, Ha M, Jung YH, et al. Low vision and the risk of dementia: a nationwide population-based cohort study. Sci Rep. 2020;10(1):9109. doi: 10.1038/s41598-020-66002-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Burton MJ, Ramke J, Marques AP, et al. The Lancet Global Health Commission on Global Eye Health: vision beyond 2020. Lancet Glob Health. 2021;9(4):e489-e551. doi: 10.1016/S2214-109X(20)30488-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Marchesi N, Fahmideh F, Boschi F, Pascale A, Barbieri A. Ocular neurodegenerative diseases: interconnection between retina and cortical areas. Cells. 2021;10(9):2394. doi: 10.3390/cells10092394 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Sonnega A, Faul JD, Ofstedal MB, Langa KM, Phillips JWR, Weir DR. Cohort profile: the Health and Retirement Study (HRS). Int J Epidemiol. 2014;43(2):576-585. doi: 10.1093/ije/dyu067 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Health and Retirement Study: 2018 HRS Core. University of Michigan. Updated September 2021. Accessed March 21, 2022. https://hrsdata.isr.umich.edu/data-products/2018-hrs-core
- 14.Flaxman AD, Wittenborn JS, Robalik T, et al. ; Vision and Eye Health Surveillance System study group . Prevalence of visual acuity loss or blindness in the US: a bayesian meta-analysis. JAMA Ophthalmol. 2021;139(7):717-723. doi: 10.1001/jamaophthalmol.2021.0527 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Bourne R, Adelson J, Flaxman S, et al. Trends in prevalence of blindness and distance and near vision impairment over 30 years: an analysis for the Global Burden of Disease Study. Lancet Glob Health. 2020;9(2):e130-e143. doi: 10.1016/S2214-109X(20)30425-3 [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
eMethods.
eTable 1. Principal-Component Analysis of Modifiable Risk Factors for Dementia in the Health and Retirement Study
eTable 2. Potentially Modifiable Dementia Risk Factors by Race and Ethnicity
eTable 3. Sensitivity Analyses for Population Attributable Fraction of Dementia Due to Vision Impairment
eFigure. Population Attributable Fraction of Potentially Modifiable Dementia Risk Factors
eReferences