Social connection, not loneliness, predicts dementia risk in Swedish older adults

Abstract

AbstractBackgroundGreater loneliness has been found to predict increased dementia risk in older adult populations. Yet, factors rather than loneliness, like objective social connections and depressive symptomatology, may explain away effects of loneliness on dementia risk. In a longitudinal sample of Swedish twins, we examined whether baseline scores and change in loneliness predict dementia risk. Additionally, we examined whether familial factors (shared genes and/or environments) confound the predictive utility of loneliness on dementia status. Finally, we hypothesized the observed effects of loneliness on dementia risk would be greater in women than men.MethodThe sample included 2,543 adult Swedish twins who completed baseline measures of loneliness and survived beyond age 60. Twins were followed until they were diagnosed with dementia, deceased, or lost to follow‐up. Analyses included first fitting latent growth curve models to evaluate change in loneliness over time; Cox regression models to test whether baseline and change in loneliness predicted dementia risk; and biometric regression models to test whether genetic and environmental pathways mediated effects of loneliness on dementia status.ResultLatent growth curve models suggest that loneliness did not change over time. Baseline loneliness did not correlate with dementia status (r = .02, t = 0.82, df = 2574, p = .413), although greater number of close friends/family predicted lower odds of dementia (Odds Ratio: 0.63, .95CI: [0.45,0.88]). As loneliness did not correlate with dementia risk), which were confirmed in biometric regression models: loneliness did not correlate with dementia via genetic or environmental pathways. The only gender difference observed was that women reporting the highest levels of loneliness were diagnosed with dementia at significantly greater rates than men reporting the highest levels of loneliness.ConclusionLoneliness did not predict dementia risk over a 30‐year follow‐up period. Despite these null findings, women reporting the highest levels of loneliness appeared to confer greater dementia risk compared to their male counterparts. Consistent with previous results, social connection (as measured by greater number of close ties) lowered dementia risk, suggesting that measures of social network density may provide greater clinical utility than perceived social isolation.