Purpose
To examine whether vision impairment and eye diseases are independently associated with memory decline in older adults.
Design
Cohort study.
Methods
We included 8,315 participants aged 50-94 years in China Health and Retirement Longitudinal Study (CHARLS) from China and 8,939 participants aged 50-95 years in Health and Retirement Study (HRS) from the United States in our analysis.
Results
During 4.0 years’ follow-up, the composite memory decreased by 0.16 points in CHARLS. During 3.9 years’ follow-up, the composite memory decreased by 0.51 in HRS. Distance vision impairment was inversely associated with an annual change in composite memory (β [95% CI]: –0.07 [–0.12, –0.01]) and immediate memory (–0.04 [–0.07, –0.02]) in CHARLS, and the corresponding values in HRS were –0.19 (–0.34, –0.05) and –0.07 (–0.13, –0.00), respectively. Near vision impairment was inversely associated with an annual change in delayed memory in CHARLS and composite memory, immediate memory, and delayed memory in HRS. In HRS, the association between distance vision impairment and memory decline was observed in individuals aged <65 years (β [95% CI]: –0.54 [–0.78, –0.30]) but not in those aged ≥65 years (–0.01 [–0.20, 0.18]). Cataract surgery or glaucoma was not significantly associated with memory decline in either CHARLS or HRS.
Conclusion
Distance vision impairment was independently associated with an accelerated rate of memory decline in both China and the United States. Near vision impairment was predictive of decline in delayed memory in China and of decline in composite, immediate, and delayed memory in the United States.
T he global number of people with dementia was 43.8 million in 2016, with 10.4 million in China and 4.0 million in the United States. The estimated worldwide cost of dementia was US$818 billion in 2015, with an increase of 35% since 2010, and the cost will be approximately US$2 trillion in 2030. The socioeconomic cost caused by dementia imposes a tremendous economic burden on both China and the United States. , The absolute number of people with dementia will still increase in the next decade owing to the increasing older population. , Therefore, modifiable determinants must be identified to prevent dementia and cognitive decline.
Both cognition and vision decrease with ageing, and vision is an important indicator of cognition in older adults. Loss of vision input may decrease activation in central sensory pathways, thus resulting in an increased risk of cognitive load, brain structure damage, and cognitive decline. , Previous studies have linked vision impairment to dementia, but the results are inconsistent. Recently, several longitudinal studies have investigated the association between vision impairment and cognitive decline; however, these studies are limited by focusing on distance vision impairment only or near vision impairment only or small sample sizes. Additionally, several studies have also examined the association of eye diseases, such as cataract and glaucoma, with cognition with inconsistent results. , Education, social support, environment, and culture may play important roles in maintaining and improving cognitive function. , Therefore, whether vision impairment and eye diseases are independently associated with cognitive decline in populations with different backgrounds and ethics remains to be explored.
The present study aimed to examine whether vision impairment and eye diseases were predictive of memory decline in China and the United States using data with representative population-based samples from each country and almost the same survey methods. We assumed that distance vision impairment, near vision impairment, cataract surgery, and glaucoma at baseline were each associated with a higher decrease in memory in both countries.
Methods
Participants
The present analysis was based on 2 nationally representative publicly available sister survey data sets: the China Health and Retirement Longitudinal Study (CHARLS) from China and the Health and Retirement Study (HRS) from the United States. ,
The China Health and Retirement Longitudinal Study
CHARLS is a nationally representative longitudinal survey of adults aged ≥45 years in China initiated in 2011 and followed up in 2013 and 2015. Multistage sampling stratified by region (urban districts and rural counties) and per capita statistics on the gross domestic product was performed to identify participants for study inclusion. In total, 150 district/county-level units were randomly selected using a probability-proportional-to-size sampling technique from a sampling frame containing all county-level units in 28 provinces across China. A face-to-face computer-assisted personal interview was conducted on 10,257 households, with a response rate of 80.5% at baseline. The participants were then followed up every 2 years.
The study protocol was approved by the Ethical Review Committee at Peking University. Written informed consent was obtained from all the participants.
The Health and Retirement Study
HRS is a nationally representative longitudinal survey of more than 37,000 individuals aged >50 years in 23,000 households in the United States. The participants were identified through a stratified, multistage, probability-sampling design, with African American and Hispanic households oversampled. Before 2004, follow-up interviews via telephone were offered to individuals aged ≤80 years and face-to-face follow-up interviews with those aged >80 years. Since 2006, half of the sample was assigned a face-to-face follow-up interview, and the other half was assigned a telephone interview. The half-samples alternated in waves, so longitudinal information from the face-to-face interview was available every 4 years at the individual level. The survey has been fielded every 2 years since 1992. The response rate for the baseline was 81.6%.
The HRS has been approved by the institutional review board at the University of Michigan. In addition, the HRS obtained informed verbal consent from the voluntary participants and followed strict procedures to protect study participants from disclosure (including maintaining a Federal Certificate of Confidentiality).
Vision Impairment and Eye Diseases
In both CHARLS and HRS, distal vision impairment was assessed by asking the participants whether their eyesight was excellent, very good, good, fair, or poor when seeing things at a distance, analogous to recognizing a friend from across the street (with glasses or corrective lenses, if applicable). Reporting poor eyesight was classified as distal vision impairment and reporting excellent, very good, good, or fair eyesight was used as the reference group. Near vision impairment was assessed by asking participants whether their eyesight was excellent, very good, good, fair, or poor when seeing things up close, such as reading ordinary newspaper print (with glasses or corrective lenses, if applicable). Reporting poor eyesight was classified as near vision impairment and reporting excellent, very good, good, or fair eyesight was used as the reference group. Cataract surgery was defined if participants reported that they ever had cataract surgery. Glaucoma was defined if participants reported that a physician ever treated them for glaucoma. In HRS, the data on vision impairment and eye diseases were collected in 1998, 2006, and 2012. Because corresponding data were collected between 2011 and 2015 in CHARLS, the data between 2012 and 2016 in HRS were used in the analysis.
Cognitive Assessment
In CHARLS, 5 cognitive screening tasks included the immediate and delayed recall of a 10-word list (one 4-word list was randomly selected), serial 7 subtraction from 100 five times, date orientation, and picture drawing. In HRS, the tests used to assess cognitive functioning included immediate and delayed recall, the serial 7 subtraction test, counting backward, object naming test, and recall of the date and president and vice president to assess orientation. The sensitivity and specificity of cognitive screening tasks used in screening for dementia in HRS was 0.78 and 0.88, respectively. Three memory tasks, including the 10-word immediate and delayed recall tests and serial 7 subtraction test of working memory that was assessed in all waves of surveys in both CHARLS and HRS, were analysed in our study.
The total score for immediate recall ranged from 0 to 10 with each correctly recalled word assigned a score of 1 and so did delayed recall. The total score for serial 7 subtraction ranged from 0 to 5 with a score of 1 assigned to each of the 5 subtractions. Composite memory scores were computed by summing the scores of all 3 memory tasks and ranged from 0 to 25 with a higher score representing better memory. Annual changes in memory scores were computed by subtracting the scores at baseline from those at follow-up divided by the follow-up duration using the following formula: annual change = (memory at follow-up – memory at baseline) / follow-up duration.
Physical Examination
In CHARLS, height was measured using a stadiometer (Seca 213; Seca Trading Co, Hangzhou, China), and weight was measured using a digital scale (Omron HN‐286; Omron Corporation, Kyoto, Japan). In HRS, height and weight were self-reported between 1992 and 2004 and then measured using a stadiometer and a digital scale since 2006. Body mass index (BMI) was computed as weight in kilograms divided by the square of height in meters.
Confounders
The data on age, gender, education (≤8 years, 9-12 years, and ≥13 years), smoking (never smoked, former smoking, and current smoking), hypertension (yes, no), diabetes (yes, no), heart disease (yes, no), stroke (yes, no), lung disease (yes, no), psychiatric disorders (yes, no), cancer (yes, no), and arthritis (yes, no) were collected using the same questions in both CHARLS and HRS. Alcohol consumption was divided into 3 groups: never, <1 drink/mo, and ≥1 drink/mo in CHARLS and <1 drink/wk, 1-6 drinks/wk, and 7 drinks/wk in HRS. In CHARLS, physical activity was assessed based on the duration and frequency spent in different physical activities, from which the metabolic equivalent of task (MET) was calculated. MET from moderate and vigorous physical activity was computed by multiplying the length (minutes per week) by an assigned MET value (4 MET for moderate physical activity and 7.5 MET for vigorous physical activity). In HRS, participants were divided into quintiles according to moderate physical activity or vigorous physical activity. Living area (rural or urban area) was also collected in CHARLS.
Statistical Analysis
The data are expressed as frequencies (percentage) and means ± standard deviations (SDs) for baseline characteristics according to country. t test was performed for continuous variables, and χ 2 test was performed for categorical variables to compare the difference in the baseline characteristics between individuals with and without distance vision impairment.
General linear regression models were used to obtain coefficients (β and 95% confidence interval [CI]) for the change in memory scores associated with vision impairment as well as eye diseases. We tested the following models: (1) age and gender; (2) model 1 plus education, smoking, alcohol intake, physical activity, and memory score (the living area was additionally adjusted in CHARLS) at baseline; and (3) model 2 plus BMI, hypertension, diabetes, heart disease, stroke, lung disease, psychiatric disorders, arthritis, and cancer at baseline. This analysis was conducted for CHARLS and HRS separately.
Moderation analysis was conducted to examine whether the association between vision impairment and memory decline was modified by important determinants for cognition, including age, gender, education, BMI, diabetes, hypertension, and psychiatric disorders.
Data analyses were conducted using SAS 9.4 for Windows (SAS Institute Inc, Cary, North Carolina, USA), and 2-sided P values <.05 were considered statistically significant.
Results
Participant Characteristics
After excluding individuals with missing data on age, sex, vision impairment, or cognition, our study sample comprised 8,315 participants aged 50.0-94.5 years in China (Supplemental Figure S1) and 8,939 participants aged 50.0-95.8 years in the United States (Supplemental Figure S2). Table 1 shows the baseline characteristics. The mean age was 61.1 ± 7.1 years (49.8% female) in CHARLS and 66.3 ± 9.4 years (50.6% female) in HRS. In both CHARLS and HRS, individuals with distance vision impairment were more likely to be older, to be less educated, to be non–alcohol drinkers, and to have a higher prevalence of hypertension, diabetes, heart disease, stroke, psychiatric disorders, lung disease, and arthritis than those without distance vision impairment (all P values <.05). Distance vision impairment was associated with lower immediate memory, delayed memory, working memory, and composite memory in both CHARLS and HRS.
| Distance Vision Impairment: China | Distance Vision Impairment: United States | |||||
|---|---|---|---|---|---|---|
| No | Yes | P Value a | No | Yes | P Value | |
| Age (y), mean ± SD | 60.7 ± 7.0 | 62.6 ± 7.4 | <.0001 | 66.2 ± 9.4 | 69.9 ± 10.0 | <.0001 |
| Gender | <.0001 | .71 | ||||
| Men | 3,442 (53.1) | 731 (39.9) | 4,184 (48.0) | 101 (49.3) | ||
| Women | 3,039 (46.9) | 1,101 (60.1) | 4,540 (52.0) | 104 (50.7) | ||
| Education | <.0001 | <.0001 | ||||
| <9 y | 4,253 (65.6) | 1,458 (79.6) | 486 (5.6) | 33 (16.1) | ||
| 9-12 y | 2,075 (32.0) | 352 (19.2) | 980 (11.2) | 37 (18.0) | ||
| >12 y | 153 (2.4) | 21 (1.1) | 7,258 (83.2) | 135 (65.9) | ||
| Missing | 2 (0.0) | 1 (0.1) | ||||
| Living area | <.0001 | |||||
| Urban | 3,752 (57.9) | 1,217 (66.4) | ||||
| Rural | 2,731 (42.1) | 615 (33.6) | ||||
| Alcohol consumption | <.0001 | |||||
| ≥1 drink/mo | 1,795 (27.7) | 392 (21.4) | ||||
| <1 drink/mo | 503 (7.8) | 113 (6.2) | ||||
| Never | 4,185 (64.6) | 1,327 (72.4) | ||||
| Alcohol consumption | .0001 | |||||
| 7 drinks/wk | 610 (7.0) | 14 (6.8) | ||||
| 1-6 drinks/wk | 3,066 (35.1) | 46 (22.4) | ||||
| <1 drink/wk | 5,032 (57.7) | 145 (70.7) | ||||
| Smoking | <.0001 | .0632 | ||||
| Current | 2,159 (33.3) | 505 (27.6) | 946 (10.8) | 33 (16.1) | ||
| Former | 627 (9.7) | 163 (8.9) | 3,741 (42.9) | 86 (42.0) | ||
| Never | 3,697 (57.0) | 1,164 (63.5) | 4,037 (46.3) | 86 (42.0) | ||
| Moderate and vigorous physical activity (MET-min/wk), mean ± SD | 1,894.7 ± 4,520.2 | 2,038.5 ± 4,622.2 | .23 | |||
| Moderate physical activity | <.0001 | |||||
| Every day | 205 (2.3) | 9 (4.4) | ||||
| >1 time/wk | 2,327 (26.7) | 35 (17.1) | ||||
| 1 time/wk | 1,095 (12.6) | 25 (12.2) | ||||
| 1-3 times/mo | 945 (10.8) | 21 (10.2) | ||||
| Never | 4,131 (47.4) | 113 (55.1) | ||||
| Missing | 21 (0.2) | 2 (1.0) | ||||
| Vigorous physical activity | <.0001 | |||||
| Every day | 698 (8.0) | 13 (6.3) | ||||
| >1 time/wk | 4,141 (47.5) | 86 (42.0) | ||||
| 1 time/wk | 1,605 (18.4) | 31 (15.1) | ||||
| 1-3 times/mo | 1,034 (11.9) | 31 (15.1) | ||||
| Never | 1,240 (14.2) | 43 (21.0) | ||||
| Missing | 6 (0.1) | 1 (0.5) | ||||
| BMI, mean ± SD | 23.51 ± 3.47 | 23.38 ± 3.49 | .15 | 29.67 ± 6.05 | 29.33 ± 5.62 | .65 |
| Hypertension | <.0001 | .0022 | ||||
| Yes | 1,634 (25.2) | 553 (30.2) | 4,775 (54.7) | 127 (62.0) | ||
| No | 4,823 (74.4) | 1,271 (69.4) | 3,934 (45.1) | 78 (38.0) | ||
| Missing | 26 (0.4) | 8 (0.4) | 15 (0.2) | |||
| Diabetes | .0004 | .0214 | ||||
| Yes | 379 (5.8) | 153 (8.4) | 1,844 (21.1) | 56 (27.3) | ||
| No | 6,053 (93.4) | 1,663 (90.8) | 6,873 (78.8) | 148 (72.2) | ||
| Missing | 51 (0.8) | 16 (0.9) | 7 (0.1) | 1 (0.5) | ||
| Heart disease | <.0001 | .0529 | ||||
| Yes | 782 (12.1) | 303 (16.5) | 1,801 (20.6) | 54 (26.3) | ||
| No | 5,672 (87.5) | 1,517 (82.8) | 6,915 (79.3) | 151 (73.7) | ||
| Missing | 29 (0.4) | 12 (0.7) | 8 (0.1) | |||
| Stroke | .0022 | .0104 | ||||
| Yes | 108 (1.7) | 52 (2.8) | 520 (6.0) | 18 (8.8) | ||
| No | 6,360 (98.1) | 1,776 (96.9) | 8,198 (94.0) | 187 (91.2) | ||
| Missing | 15 (0.2) | 4 (0.2) | 6 (0.1) | |||
| Psychiatric disorders | .0446 | .0314 | ||||
| Yes | 38 (0.6) | 21 (1.1) | 1,207 (13.8) | 37 (18.0) | ||
| No | 6,416 (99.0) | 1,803 (98.4) | 7,507 (86.0) | 168 (82.0) | ||
| Missing | 29 (0.4) | 8 (0.4) | 10 (0.1) | |||
| Lung disease | <.0001 | .0142 | ||||
| Yes | 628 (9.7) | 259 (14.1) | 624 (7.2) | 23 (11.2) | ||
| No | 5,833 (90.0) | 1,566 (85.5) | 8,094 (92.8) | 181 (88.3) | ||
| Missing | 22 (0.3) | 7 (0.4) | 6 (0.1) | 1 (0.5) | ||
| Arthritis | <.0001 | .0011 | ||||
| Yes | 2,056 (31.7) | 854 (46.6) | 4,690 (53.8) | 129 (62.9) | ||
| No | 4,417 (68.1) | 972 (53.1) | 4,026 (46.1) | 76 (37.1) | ||
| Missing | 10 (0.2) | 6 (0.3) | 8 (0.1) | |||
| Cancer | .10 | .51 | ||||
| Yes | 53 (0.8) | 21 (1.1) | 1,149 (13.2) | 30 (14.6) | ||
| No | 6,402 (98.8) | 1,806 (98.6) | 7,560 (86.7) | 175 (85.4) | ||
| Missing | 28 (0.4) | 5 (0.3) | 15 (0.2) | |||
| Immediate memory, mean ± SD | 4.19 ± 1.60 | 3.71 ± 1.51 | <.0001 | 5.59 ± 1.57 | 4.94 ± 1.84 | <.0001 |
| Delayed memory, mean ± SD | 3.23 ± 1.89 | 2.74 ± 1.80 | <.0001 | 4.61 ± 1.89 | 3.87 ± 2.13 | <.0001 |
| Working memory, mean ± SD | 3.09 ± 1.93 | 2.52 ± 1.97 | <.0001 | 3.69 ± 1.55 | 3.19 ± 1.86 | <.0001 |
| Composite memory, mean ± SD | 10.52 ± 4.13 | 8.97 ± 4.02 | <.0001 | 13.90 ± 3.94 | 12.00 ± 4.45 | <.0001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
