The Association of Cognitive and Visual Function in a Nationally Representative Study of Older Adults in India

Abstract

Introduction:

Due to population aging, India is poised to experience a large increase in the burden of both dementia and vision impairment (VI). Prior studies from other settings suggest that VI may be a modifiable risk factor for cognitive decline and dementia. However, to date, no studies have examined the association of impaired visual acuity and cognition in India.

Methods:

A total of 3,784 participants in wave 1 of the population-based Diagnostic Assessment of Dementia for the Longitudinal Aging Study in India underwent visual acuity testing and a cognitive test battery. Multivariable linear regression was used to model the cross-sectional associations of mild (<6/12–6/16), moderate (<6/18–6/60), and severe visual acuity impairment/blindness (<6/60) with cognitive performance scores corresponding to total cognition, orientation, memory, language/fluency, executive function, and informant reported cognitive status. Models were adjusted for demographic, socioeconomic, and health characteristics.

Results:

The weighted percentage of participants with any VI was 52.6%. Vision impairment was independently associated with lower cognitive scores across all domains, even after adjustment for known dementia risk factors. In fully adjusted models of total cognition (mean score: 130.7), mild, moderate, and severe VI/blindness were associated with a significant change of −3.5 (95%CI: −6.3, −0.6), −8.2 (95%CI −10.5, −5.6), and −16.8 (95%CI −22.3, −11.3) units, respectively. A dose-response association between level of VI and cognitive function was observed for all cognitive outcomes except for language/fluency domain scores. Associations were robust when cognitive tests dependent on visual function were excluded from outcomes. Across each fully-adjusted model of total, domain-specific, and informant-reported cognitive performance, moderate VI was equivalent to 5–9 years of cognitive aging.

Discussion/Conclusion:

This study illustrates that VI is cross-sectionally associated with lower cognitive performance, largely in a dose-response pattern, across various cognitive domains in the Indian population. These findings are important for informing future longitudinal and interventional studies.

Introduction

Dementia is a leading cause of disability and death globally [1]. In 2016, an estimated 43–47 million people were living with dementia with an expected increase to over 100 million by 2050. The largest increase is projected in low- and middle-income countries (LMICs) due to a large aging population and increased longevity [1, 2]. In India, the over 60 population is projected to increase from 8% in 2015 to 19% in 2050, accounting for approximately 320 million people [3]. The number of adults with dementia in India is projected to increase concomitantly over this period [4].

Prior studies have reported that up to 41% of the global burden of dementia may be due to known, potentially modifiable risk factors [1, 5, 6]. Vision impairment (VI), which affects over 1 billion people globally [7], may be one such modifiable risk factor since up to 80% of VI and blindness is avoidable or treatable, often with low-cost interventions like cataract surgery and eyeglasses [8]. In various settings and populations, prior studies have demonstrated that VI is associated with cognitive decline and dementia [9–14]. However, several important gaps remain in understanding the nature of this association. For example, while various hypotheses have been proposed to account for the association between VI and cognitive decline and dementia, to date little research has been done test hypothesized pathways [15].

There is also a paucity of research on the association between visual and cognitive function in LMICs. Data from high-income countries may not be readily generalizable to LMICs due to the role of culture and geographic context in shaping health, disease, and the aging process [16]. Additionally, access to health services, ongoing epidemiologic transitions, and variation in the causes of vision loss and cognitive decline between settings may impact the vision-cognition relationship. A recently published systematic review identified only eight published studies that have examined this association in LMICs globally and the only one from India relied on self-reported, rather than objectively measured, visual function [17]. Moreover, only one of these eight studies reported associations with VI and specific domains of cognitive functioning. Thus, there remain critical gaps in understanding the relationships between visual and cognitive function in LMICs, including in India, which is home to the greatest number of individuals with blindness and VI in the world [7].

The Diagnostic Assessment of Dementia for the Longitudinal Aging Study in India (LASI-DAD) is a population-based study that represents the entire Indian population aged 60 and older. Through its focus on late-life cognition, LASI-DAD provides novel opportunities to study the association of cognition with economic, demographic, and health characteristics of older Indian adults. In this study we examined the association of VI with overall and domain-specific cognitive function in LASI-DAD.

Materials and Methods

Data Source

The Longitudinal Aging Study in India (LASI) consists of over 72,000 community-dwelling adults age 45 and older and their spouses and is representative of the country, as well as each state and union territory in India. The study is harmonized with the Health and Retirement Study (HRS) in the United States and other HRS sister-studies worldwide. Several studies in the HRS network have launched separate in depth studies of late-life cognition and dementia using a harmonized cognitive assessment protocol (HCAP) that is comparable across geographic and cultural contexts [18]. The LASI-DAD, one of the HCAP studies, consists of 4,096 participants age 60 years and older from 18 states and union territories that are representative of the Indian population. Participants were recruited into LASI-DAD through a stratified random sampling process that ensured an approximately equivalent number of participants at high and low risk of cognitive impairment based on results of cognitive test results in LASI. The LASI-DAD study protocol has been described in detail elsewhere [18, 19].

Cognitive Performance Measures

The LASI-DAD cognitive test protocol was developed in consideration of illiteracy and innumeracy in the population. Cognitive function was measured using a summary total cognition variable, as well as domain-specific summary variables. The total cognition variable was developed to assess overall cognitive function and has been described previously [19]. A variant of this total cognition variable was also constructed that omitted cognitive tests that were largely dependent on vision. Domain-specific cognitive performance was assessed using summary variables based on the factor structure of the LASI-DAD cognitive battery described by Gross et al [20]. The Supplementary Appendix lists the individual cognitive tests that comprised each of these summary variables. Additionally, informant assessment of cognition was modeled using results of the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) [21].

Vision Impairment

Vision was measured in the LASI survey using a Tumbling E chart displayed on a laptop situated at 3 meters distance from the participant. Each eye was tested separately using the participant’s habitual refractive correction (e.g., eyeglasses, contact lenses) for distance vision, if available. Categorization of VI was based on World Health Organization definitions using measured visual acuity in the better-seeing eye [22]: mild VI (<6/12–6/18), moderate VI (<6/18–6/60), severe VI (<6/60–3/60), and blindness (<3/60). Due to the relatively small number of participants with severe VI and blindness, these were collapsed into a single category for analyses.

Covariates

We considered conceptually relevant co-variates based on existing literature on known risk factors for VI and for dementia [6, 8, 23, 24]. The following socioeconomic and demographic covariates were included: age, sex, highest education (illiterate, primary school, secondary school or more), marital status (married, widowed, other), urbanicity (urban, rural), and per capita household consumption quartile. We also considered the following health-related variables: self-reported diagnosis of diabetes, heart disease, hypertension, and stroke; a history of having ever been a smoker; and body-mass index (BMI). Body-mass index was calculated based on weight and height measured by the LASI-DAD research team.

Statistical Analyses

We first conducted descriptive analyses to report the prevalence of VI and to examine bivariate associations between VI and demographic variables, socioeconomic status, and health measures. We accounted for survey design features and applied survey weights. We formally tested the differences between VI status and covariates, using the Pearson’s chi-square test. For all cognitive test variables there was complete data for ≥85% of the sample. Thus, observations that were missing, including due to failure to complete a task because of illiteracy or innumeracy, were imputed, as is common in survey data analysis; the imputation method for cognitive test data in LASI-DAD has been previously described [20]. We then conducted multivariable ordinary least squares regression analyses to examine the association between cognitive measures and VI status, while adjusting for other covariates. For each cognitive outcome, we conducted four separate analyses. The first model only included VI status without adjusting any covariates; the second was adjusted for age and sex; the third added demographic and socioeconomic covariates (education, marital status, urbanicity, and consumption); and the fourth was fully-adjusted and contained all aforementioned covariates in addition to health-related variables (BMI, smoking status, and self-reported diagnoses of diabetes, stroke, hypertension, and heart disease). All analyses were conducted using Stata version 14.2. Cognitive aging refers to changes in cognitive performance that occur as a part of the aging process [25]. To calculate the number of years of cognitive aging with which VI was equivalent, we assumed that the change in cognition from age 60–69 to age 70–79 represented 10 years of cognitive aging.

Results

The characteristics of the analytic sample (n = 3,784), stratified by VI status, are presented in Table 1. Overall, 52.6% had VI; specifically, 16.5%, 32.4%, and 3.7% had mild VI, moderate VI, and severe VI/blindness, respectively. Those with VI were more likely to be older, have lower educational attainment, be unmarried, live in a rural setting, belong to a lower consumption quartile, have a lower BMI, and be non-diabetic; they were no more likely to be female (p = 0.12), or have heart disease (p = 0.29), hypertension (p = 0.49), or a history of stroke (p = 0.79) or smoking (p = 0.61).

Table 1.

Weighted sample characteristics by vision impairment status.

			Vision Impairment Status (%)a				P-valueb
Characteristics	Total (n)	Total (%)c	$\frac{\text{Normal}}{\ge \mathrm{6/12}}$	$\frac{\text{Mild VI}}{<\mathrm{6/12-6/18}}$	$\frac{\text{Moderate VI}}{<\mathrm{6/18-6/60}}$	$\frac{\text{Severe VI/Blind}}{<\mathrm{6/60}}$
Total, % (n = 3784)			47.4	16.5	32.4	3.7
Age							<.001
60–69	2,312	62.5	54.1	16.5	26.7	2.6
70–79	1,090	27.7	39.5	18.0	38.3	4.1
80+	382	9.8	27.1	11.8	51.6	9.5
Gender							0.12
Male	1,745	49.2	49.0	15.1	32.6	3.2
Female	2,039	50.8	45.9	17.8	32.1	4.3
Education							<.001
None	1,852	55.5	41.5	17.2	36.2	5.1
Primary School	1,003	23.7	47.9	16.6	33.3	2.2
≥ Secondary School	929	20.8	62.8	14.4	21.0	1.8
Marital Status							<.001
Married	2,491	65.7	49.9	17.2	30.1	2.8
Widowed	1,224	32.7	43.6	15.1	36.0	5.4
Other	69	1.7	24.9	16.3	51.0	7.9
Urbanicity							<.001
Urban	1,434	29.0	55.5	16.6	25.7	2.2
Rural	2,350	71.0	44.1	16.4	35.1	4.4
Consumption Quartile							<.001
1st	947	30.7	43.6	16.6	33.6	6.2
2nd	946	24.9	46.2	16.4	34.8	2.6
3rd	946	23.3	48.3	17.0	32.0	2.7
4th	945	21.1	53.5	15.8	28.2	2.5
BMI (kg/m2)							<.001
<18.5	850	26.1	37.0	17.6	39.7	5.7
18.5–24.9	1,837	49.9	49.7	15.7	31.2	3.4
25.0–29.9	762	17.7	55.9	16.2	26.6	1.3
≥30.0	287	6.3	51.6	19.5	25.8	3.1
Diabetes							.01
No	3,115	86.2	46.2	16.7	33.4	3.6
Yes	669	13.8	54.9	14.8	25.8	4.5
Heart Disease							.29
No	3,526	94.7	46.9	16.5	32.7	3.8
Yes	258	5.3	55.9	16.5	25.6	2.0
Hypertension							.49
No	2,298	66.1	47.1	16.4	32.8	3.8
Yes	1,486	34.0	48.1	16.6	31.5	3.7
Stroke							.79
No	3,680	97.3	47.4	16.6	32.4	3.7
Yes	104	2.7	48.3	13.9	32.5	5.4
Ever smoked							.61
No	2,949	78.7	47.3	16.6	32.1	4.0
Yes	835	21.3	47.9	16.0	33.3	2.8

BMI: body-mass index; VI: Vision Impairment

^abased on visual acuity in the better-seeing eye;

^bPearson chi-squared test;

^cTable contains raw counts and survey-weighted percentages, so percentages may not sum to 100%

Table 2 presents summary statistics for each of the cognitive performance outcomes, both overall and stratified by VI status. These unadjusted scores were highest in respondents with no VI and each shows a step-wise decrease across categories of worsening VI. The results from linear models of total cognition are presented in Table 3. Total cognition scores were significantly associated with VI in a step-wise pattern, wherein worse VI was associated with lower cognitive performance scores, across all regression models. When cognitive tests that depended strongly on vision were removed from the total cognition variable, the association between VI and cognition was similar across all models. In the fully-adjusted total cognition models, moderate VI was equivalent to 7.0–8.5 years of cognitive aging.

Table 2.

Summary statistics for outcome variables.

Outcome	Overall (n=3,784)			Vision Impairment Status
				No VI	Mild VI	Moderate VI	Severe VI/Blind
	Mean	SD	Range	Mean
Total Cognition 1a	130.7	46.0	7–282	136.8	124.9	114.2	93.3
Total Cognition 2b	102.3	33.5	7–214	105.7	97.8	91.1	78.3
Orientation	9.7	2.8	0–13	10.1	9.4	9.0	8.0
Memory	58.2	21.8	1–147	60.1	54.7	51.1	43.9
Language/Fluency	19.5	5.4	0–42	19.8	19.1	18.3	17.4
Executive Function	17.0	9.3	0–32	17.8	16.5	14.6	10.5
IQCODE	3.4	0.5	1–5	3.3	3.4	3.5	3.7

IQCODE: Informant Questionnaire on Cognitive Decline in the Elderly; SD: standard deviation, VI: vision impairment

^aTotal cognition 1 includes all tests in the cognitive test battery

^bTotal cognition 2 excludes cognitive tests that were largely dependent on vision (see Supplementary Appendix)

Table 3.

Association of vision impairment and total cognition.

Domain	Model 1a β (95% CI)	Model 2b β (95% CI)	Model 3c β (95% CI)	Model 4d β (95% CI)
Total Cognition 1e
No VI	Reference	Reference	Reference	Reference
Mild VI	−11.9 (−15.8, −8.0)**	−9.3 (−13.1, −5.6)**	−4.1 (−6.9, −1.2)**	−3.5 (−6.3, −0.6)*
Moderate VI	−22.6 (−25.7, −19.5)**	−18.0 (−21.0, −15.0)**	−8.9 (−11.3, −6.5)**	−8.2 (−10.5, −5.8)**
Severe VI/Blind	−43.5 (−50.8, −36.1)**	−34.9 (−41.9, −27.9)**	−20.0 (−25.5, −14.5)**	−16.8 (−22.3, −11.3)**
Total Cognition 2f
No VI	Reference	Reference	Reference	Reference
Mild VI	−7.9 (−10.8, −5.1)**	−6.2 (−9.0, −3.5)**	−2.7 (−5.0, −0.5)*	−2.1 (−4.4, 0.1)
Moderate VI	−14.6 (−16.9, −12.3)**	−11.4 (−13.6, −9.2)**	−5.4 (−7.2, −3.5)**	−4.7 (−6.6, −2.9)**
Severe VI/Blind	−27.4 (−32.8, −22.1)**	−21.6 (−26.8, −16.4)**	−11.5 (−15.9, −7.2)**	−9.0 (−13.3, −4.7)**

CI: confidence interval, VI: vision impairment

^aModel 1: unadjusted

^bModel 2: age and sex adjusted

^cModel 3: Model 2 and education, marital status, urbanicity, and consumption quartile

^dModel 4: Model 3 and body-mass index, diabetes, heart disease, hypertension, stroke, and smoking status

^eTotal cognition 1 includes all cognitive tests

^fTotal cognition 2 excludes any cognitive tests that largely dependent on vision

^*p<.05

^**p<.01

Domain-specific cognitive performance was also modeled. Across all cognitive domains, VI was independently associated with lower cognitive performance scores (Table 4). There was a clear step-wise pattern between worse VI and worse cognitive performance in the orientation, memory, and executive function domains across all levels of model adjustment. Informant-reported cognitive function also followed this pattern. However, scores on the language/fluency domain, while significantly lower in those with VI compared to normal vision, did now follow a step-wise pattern after adjustment for demographic (model 3) and health factors (model 4). In fully-adjusted models, moderate VI was equivalent to 9 years of cognitive aging in the orientation domain, 7 years in memory, 9 years in language/fluency, 5 years in executive function, and 7 years in informant reported cognitive function.

Table 4.

Association of vision impairment and domain-specific cognitive function.^a

	Model 1b β (95% CI)	Model 2c β (95% CI)	Model 3d β (95% CI)	Model 4e β (95% CI)
Orientation
No VI	Reference	Reference	Reference	Reference
Mild VI	−0.7 (−0.9, −0.4)**	−0.5 (−0.7, −0.3)**	−0.2 (−0.4, −0.0)*	−0.2 (−0.4, −0.0)*
Moderate VI	−1.1 (−1.3, −0.9)**	−0.9 (−1.0, −0.7)**	−0.4 (−0.6, −0.3)**	−0.4 (−0.5, −0.2)**
Severe VI/Blind	−2.0 (−2.5, −1.6)**	−1.6 (−2.0, −1.1)**	−0.7 (−1.1, −0.4)**	−0.7 (−1.1, −0.3)**
Memory
No VI	Reference	Reference	Reference	Reference
Mild VI	−5.3 (−7.2, −3.5)**	−4.5 (−6.3, −2.7)**	−2.5 (−4.1, −0.9)**	−2.2 (−3.8, −0.6)**
Moderate VI	−9.0 (−10.5, −7.5)**	−7.1 (−8.5, −5.6)**	−3.6 (−4.9, −2.3)**	−3.2 (−4.5, −2.0)**
Severe VI/Blind	−16.2 (−19.6, −12.7)**	−12.8 (−16.2, −9.4)**	−7.2 (−10.2, −4.2)**	−5.7 (−8.7, −2.6)**
Language/Fluency
No VI	Reference	Reference	Reference	Reference
Mild VI	−0.7 (−1.2, −0.2)**	−0.5 (−1.0, −0.1)*	−0.1 (−0.6, 0.3)	−0.1 (−0.6, 0.3)
Moderate VI	−1.5 (−1.9, −1.1)**	−1.2 (−1.5, −0.8)**	−0.5 (−0.9, −0.2)**	−0.5 (−0.8, −0.1)*
Severe VI/Blind	−2.4 (−3.3, −1.5)**	−1.8 (−2.7, −0.9)**	−0.8 (−1.7, 0.0)	−0.5 (−1.3, 0.3)
Executive Function
No VI	Reference	Reference	Reference	Reference
Mild VI	−1.4 (−2.2, −0.5)**	−0.8 (−1.6, −0.1)*	0.1 (−0.5, 0.8)	0.3 (−0.3, 0.9)
Moderate VI	−3.3 (−3.9, −2.6)**	−2.5 (−3.1, −1.9)**	−0.9 (−1.4, −0.4)**	−0.7 (−1.3, −0.2)**
Severe VI/Blind	−7.3 (−8.8, −5.8)**	−5.8 (−7.2, −4.3)**	−3.0 (−4.2, −1.8)**	−2.3 (−3.6, −1.1)**
IQCODE
No VI	Reference	Reference	Reference	Reference
Mild VI	0.1 (0.1, 0.1)**	0.1 (0.0, 0.1)**	0.1 (0.0, 0.1)*	0.1 (0.0, 0.1)*
Moderate VI	0.2 (0.1, 0.2)**	0.1 (0.1, 0.2)**	0.1 (0.0, 0.1)**	0.1 (0.0, 0.1)**
Severe VI/Blind	0.3 (0.2, 0.4)**	0.3 (0.2, 0.4)**	0.2 (0.1, 0.3)**	0.2 (0.1, 0.2)**

CI: confidence interval, IQCODE: Informant Questionnaire on Cognitive Decline in the Elderly, VI: vision impairment

^aFor all outcomes lower scores represent worse cognitive function, except for the IQCODE where high scores represent greater cognitive decline

^bModel 1: unadjusted

^cModel 2: age and sex adjusted

^dModel 3: Model 2 and education, marital status, urbanicity, and consumption quartile

^eModel 4: Model 3 and body-mass index, diabetes, heart disease, hypertension, stroke, and smoking status

^*p<.05

^**p<.01

Discussion/Conclusion

In a nationally-representative cohort of older Indian adults, this study found that VI was strongly and independently associated with poorer cognitive performance, largely in a dose-response pattern. To our knowledge, this is the first study to investigate the association of objectively measured VI with cognitive outcomes in the Indian population. Moreover, this study provides novel evidence for the association of VI with specific domains of cognitive function, including orientation, memory, language/fluency, and executive function.

There is an urgent need to understand more fully the associations between visual and cognitive function in India. Due to ongoing demographic transitions, nearly 1 in 5 Indians is projected to be age 60 years or older by 2050, compared to just 1 in 12 in 2015 [3], representing a massive increase in the population at highest risk for cognitive impairment and dementia. Additionally, India is home to nearly one-quarter of the blind and visually impaired people in the world, approximately 80% of whom are 50 years or older [7]. Since age is a strong common risk factor for both vision and cognitive impairments, these data suggest that India is likely poised to experience a large increase in the number of older adults with co-occurring visual and cognitive pathology over the coming decades.

Blindness and VI could represent a key modifiable risk factor for cognitive impairment and dementia in India and elsewhere. An estimated 80% of VI and blindness is preventable or curable [8] and more than half of cases are attributable to cataract or uncorrected refractive error [26], both of which can be cured with highly-cost effective interventions. However, ultimately, large-scale randomized trials are needed to determine whether intervening to improve visual function can delay or slow cognitive decline and the risk of incident dementia. First, however, population-based data like those presented herein are vital for characterizing this association. While wave 1 of LASI-DAD provides nationally-representative data on the cross-sectional association, future waves will contribute to a longitudinal understanding.

Prior studies, largely from high-income countries, have reported a consistent association between vision and cognitive outcomes [9–13]. In a forthcoming systematic review, we identified 110 studies on the association of VI and cognitive outcomes, 91 (83%) of which reported a positive association. One study from the United States showed that while VI strongly predicted future cognitive decline, cognitive performance was not a strong predictor of visual outcomes [9]. In a study from Singapore in which 31% of participants were of Indian ancestry, baseline vision and visual decline over 6 years were both strongly predictive of accelerated cognitive decline compared to those with better vision and slower visual decline [13].

A recent systematic review identified only eight published studies on the association of vision with cognition or dementia in LMICs [17]. Of these, five employed objective measures of visual function (e.g., an eye chart), two relied on self-reported visual difficulty, and one did not report its methodology. The only study from India, by the 10/66 Dementia Research Group, found that self-reported visual difficulty was not consistently associated with dementia or with cognition among those who had dementia in adjusted models [27]. However, among the five studies from a diverse group of Asian and African LMICs that analyzed objectively measured visual function, there was a consistent significant association between worse visual and cognitive function after adjustment for likely confounders [28–32]. To date, no similar studies have been conducted in India.

In the current study, we found that VI and cognitive performance were associated in a dose-response pattern in the LASI-DAD sample, which is weighted to be representative of the entire over 60 Indian population. Specifically, worse categories of VI were associated with progressively lower total cognition, orientation, memory, executive function, and informant-reported cognition, even after adjustment for known demographic and socioeconomic associations with cognitive impairment, dementia, and VI [6, 23]. This pattern was not observed in fully-adjusted models of language/fluency. The reasons for differential domain-specific effects are not clear, but possibilities include the specific effects of VI on brain structure and function or the effect of dual-tasking on specific cognitive processes.

Recently published studies from the Women’s Health Initiative (WHI) in the United States [12] and from a large cohort of older adults in China [14] reported a similar dose-response relationship between worse categories of VI and incident dementia. In the WHI, over a mean 3.8 years of follow-up, the hazard ratio for incident dementia among participants with baseline visual acuity <6/12, <6/24, and <6/30 was 2.1, 5.2, and 5.7, respectively, with similar findings in models of incident mild cognitive impairment [12]. In the Chinese study, over 15,000 older adults were followed for 6 years and mild, moderate, and severe VI at baseline were associated with a hazard ratio for incident dementia of 1.2, 2.1, and 8.7, respectively. The consistent dose-response relationship in these studies and in LASI-DAD contributes to our understanding of the association between visual and cognitive function, though considerable additional work is needed, particularly in LMICs like India.

Findings from the current study may have implications for future epidemiological and interventional research in India. Few prior studies have examined the association between vision and cognitive function across multiple domains [33, 34]. In a cohort from the U.S., Dearborn et al reported that VI was associated with future declines in visuospatial, verbal episodic memory, and executive function domains, but not working memory or scanning and tracking [34]. While that study reported associations over 5 years of follow-up, data were not analyzed to determine whether associations varied based on VI severity. Future research in India is needed to characterize longitudinal domain-specific associations, which may help to inform the choice of outcome measures and interventions in future trials that aim to slow cognitive decline through optimizing vision.

There were several unexpected findings in this study. First, those with diabetes were less likely to have VI, though diabetes is an important cause of VI and blindness in India and globally [35]. This may have been the case if those who self-reported a diagnosis of diabetes were also more likely to have received routine medical care, including eye exams and vision enhancing interventions like eyeglasses and cataract surgery. Additionally, low BMI (<18.5 kg/m²) was associated with both VI and lower cognitive performance, while high BMI (>25.0 kg/m²) was associated with better cognitive performance. Previous studies in India have shown that obesity is more common among those with higher education and socioeconomic status [36], which may help to account for this phenomenon. Moreover, undernutrition is associated with a diverse set of adverse health outcomes (including VI that can be cause by vitamin deficiencies), and may contribute to this finding [37]. Compared to current data from the Global Burden of Disease (GBD) project, data from LASI-DAD estimate a higher prevalence of mild and moderate VI and a similar prevalence of blindness in India. Of note, however, there are no up-to-date national-level data from India included in the GBD models, so data from LASI and LASI-DAD may make an important contribution to understanding the current epidemiology of VI and blindness at the state and national-level in India [7].

This study had several limitations. First, data were cross-sectional so it was not possible to assess the directionality of associations or to prospectively assess the association between changes in vision and cognition. Second, it is possible that participants with VI had difficulty seeing to complete vision-dependent cognitive tests, which could have affected performance, though our findings were robust in a model that excluded cognitive tests that depended strongly on vision. Likewise, it is plausible that cognitive impairment could impact visual perception, resulting in a cross-sectional association due to reverse causality. Duration and cause of vision loss could impact cognitive function, though these data were not available in LASI or LASI-DAD. Due to the nature of observational data, there is also a possibility of residual confounding and that cognitive effects attributable to vision may be reflective of differences in demographics, though we sought to control for key demographic and socioeconomic variables. This study also had a number of strengths. As a population-based study, findings from LASI-DAD are nationally-representative. Future waves of LASI-DAD will provide longitudinal data that will be used to build on this study’s findings in order to describe the longitudinal association of vision and vision changes with cognitive trajectories and incident dementia. Findings from the IQCODE, a validated informant-reported measure of cognitive status, corroborate the results of performance tests, which may be particularly relevant given the high-degree of demographic heterogeneity in the sample. Additionally, LASI and LASI-DAD are part of the HRS and HCAP families of studies, which facilitates cross-national comparisons of risk factors for late life cognitive decline and dementia.

In conclusion, this study used novel data from LASI-DAD to describe the association of VI with cognitive performance across multiple cognitive domains in the Indian population. As in other geographic contexts, VI appears to be strongly and independently associated with worse cognitive outcomes in India. Given the large and growing population of older adults in India, there is a pressing need to characterize potentially modifiable risk factors for cognitive decline and dementia in this population. The current study provides foundational epidemiologic data for future longitudinal and interventional studies with the overarching goal of characterizing and decreasing the population burden of cognitive decline and dementia in India.

Fig. 1. Association between severity of vision impairment and total cognition scores.

The figure illustrates that worse categories on vision impairment were associated with lower total cognition scores.

VI: vision impairment

*Total cognition 1 includes all tests in the cognitive test battery

^†Total cognition 2 excludes any cognitive tests that were strongly dependent on vision