The Relationship Between Cognitive Decline and All-Cause Mortality Is Modified by Living Alone and a Small Social Network: A Paradox of Isolation.

Abstract

Although cognitive decline is a well-known mortality risk, it has not been adequately investigated, whether social relationships modify the relationship between cognitive decline and mortality. We examined the modifying effects of social relationships (household composition, social network [frequency of social contact with individuals outside the household], and social participation) on the association between cognitive decline and all-cause mortality in older Japanese people. In 2015, a baseline questionnaire was distributed to all 132,005 independent community-dwelling individuals aged ≥65 years resident in Adachi Ward of the Tokyo Metropolitan area. The final sample analyzed comprised 74,872 participants (men: 44.9%; mean age: 73.7 ± 6.0 years). Cognitive decline was assessed using a self-administered dementia checklist that was validated using the Clinical Dementia Rating Scale. A Cox proportional hazard model with an average follow-up of 1,657 days revealed that cognitive decline was associated with higher mortality (hazard ratio [HR]: 1.37, 95% confidence interval [95% CI]: 1.25-1.50). We identified significant associations among household composition, social networks, and cognitive decline. Stratified analyses indicated that the cognitive decline-mortality association was stronger among participants with low contact frequency (HR = 1.60, 95% CI: 1.39-1.85) than high frequency (HR = 1.24, 95% CI: 1.11-1.39). Conversely, the association was weaker among individuals living alone (HR = 1.13, 95% CI: 0.90-1.40) than among cohabiting individuals (HR = 1.43, 95% CI: 1.29-1.57). Although living alone and having a small social network represent an isolated status, their modifying effects were the opposite. These findings indicate that the isolation type should be considered when implementing support strategies for older adults with cognitive decline or dementia.