Introduction:
Recent evidence supports relationships between social engagement and cognitive function, but the importance of social engagement decades before the onset of cognitive impairment is less well described. We examined associations of midlife social isolation and social support with mid-to-late life cognitive decline.
Methods:
In 14,083 participants in the Atherosclerosis Risk in Communities (ARIC) Study, social isolation was evaluated in 1990-92 by the Lubben Social Network Scale, discretized into: Isolated, and high, moderate, and low “risk” for isolation. Perceived social support, using the Interpersonal Support Evaluation List-Short Form, was discretized into tertiles. Linear mixed effects models with random intercepts and slopes were used to estimate the cross-sectional and longitudinal (1990-92 to 2011-13) associations of social isolation/social support with general cognitive performance. Models were adjusted for demographic factors and comorbidities, with effect modification explored by race and sex. We used multiple imputation by chained equations (MICE) to account for attrition.
Results:
Participants were on average 57.0 years old, 56% female, and 24% Black; 3% were “isolated” and 38% reported low social support. On average, men had a higher level of social isolation and lower level of social support. White participants had lower scored risk of social isolation. Socially isolated participants had on average a 0.253 standard deviation (SD) (95% CI 0.181, 0.324) lower baseline cognitive score, but a similar rate of cognitive decline in both the first 6 (difference: 0.002 SD/year, 95% CI -0.008, 0.012) and subsequent 15 years (difference: 0.001 SD/year, 95% CI: -0.006, 0.008), compared to those with lower social isolation level. Compared to participants with the highest level of social support, participants with the lowest level of social support had on average a 0.183 SD (95% CI 0.154, 0.212) lower baseline cognitive score, with a similar rate of cognitive decline in the first 6 years (difference: -0.001 SD/year, 95% CI -0.005, 0.003), but a 0.005 SD/year (95% CI 0.002, 0.007) slower rate of cognitive decline after 6 years, mainly driven by white and male participants. MICE-imputed results did not change.
Conclusions:
Although higher levels of social isolation and lower levels of social support in midlife were associated with worse cognitive function at baseline, these exposures had only a minimal impact on 20-year cognitive change. Some relationships differed by race and sex. Further studies are needed to better elucidate whether interventions addressing social isolation and support would be effective in the prevention of cognitive decline and dementia.
Social isolation, rather than loneliness, is associated with cognitive decline in older adults: the China Health and Retirement Longitudinal Study