Association of Vision Loss With Cognition in Older Adults

IMPORTANCE Visual dysfunction and poor cognition are highly prevalent among older adults; however, the relationship is not well defined. OBJECTIVE To evaluate the association of measured and self-reported visual impairment (VI) with cognition in older US adults. DESIGN, SETTING, AND PARTICIPANTS Cross-sectional analysis of 2 national data sets: the National Health and Nutrition Examination Survey (NHANES), 1999-2002, and the National Health and Aging Trends Study (NHATS), 2011-2015. The NHANES was composed of a civilian, noninstitutionalized community, and the NHATS comprised Medicare beneficiaries in the contiguous United States. Vision was measured at distance, near, and by self-report in the NHANES and by self-report alone in the NHATS. Sample weights were used to ensure result generalizability. MAIN OUTCOMES AND MEASURES The NHANES measured Digit Symbol Substitution Test (DSST) score and relative DSST impairment (DSST score <= 28, lowest quartile in study cohort), and the NHATS measured probable or possible dementia, classified per NHATS protocol. RESULTS The NHANES comprised 2975 respondents aged 60 years and older who completed the DSST measuring cognitive performance. Mean (SD) age was 72 (8) years, 52% of participants were women (n = 1527), and 61% were non-Hispanic white (n = 1818). The NHATS included 30 202 respondents aged 65 years and older with dementia status assessment. The largest proportion (40%; n = 12 212) were between 75 and 84 years of age. Fifty-eight percent were women (n = 17 659), and 69% were non-Hispanic white (n = 20 842). In the NHANES, distance VI (beta = -5.1; 95% CI, -8.6 to -1.6; odds ratio [OR], 2.8; 95% CI, 1.1-6.7) and subjective VI (beta = -5.3; 95% CI, -8.0 to -2.6; OR, 2.7; 95% CI, 1.6-4.8) were both associated with lower DSST scores and higher odds of DSST impairment after full adjustment with covariates. Near VI was associated with lower DSST scores but not higher odds of DSST impairment. The NHATS data corroborated these results, with all vision variables associated with higher odds of dementia after full adjustment (distance VI: OR, 1.9; 95% CI, 1.6-2.2; near VI: OR, 2.6; 95% CI, 2.2-3.1; either distance or near VI: OR, 2.1; 95% CI, 1.8-2.4). CONCLUSIONS AND RELEVANCE In a nationally representative sample of older US adults, vision dysfunction at distance and based on self-reports was associated with poor cognitive function. This was substantiated by a representative sample of US Medicare beneficiaries using self-reported visual function, reinforcing the value of identifying patients with visual compromise. Further study of longitudinal interactions between vision and cognition is warranted.