Social Networks and Incident Alzheimer’s Disease: the Kuakini Honolulu-Asia Aging Study

Background: We assessed the longitudinal association between social networks in late life and the incidence of all-cause dementia (DEM), Alzheimer’s disease (AD) and vascular dementia (VaD) over a 10-year follow-up period in a community-based cohort of older Japanese-American men. Method: This prospective study analyzed data from cognitively intact participants of the Kuakini Honolulu-Asia Aging Study (K-HAAS). The K-HAAS began with the Kuakini Honolulu Heart Program’s fourth examination in 1991-1993, which served as our baseline timepoint. We followed 2,636 Japanese-American men through the seventh exam (1999-2000). Participants were dementia-free at baseline and during the first 3 years of follow-up. Social network strength was assessed by the Lubben Social Network Scale (LSNS). Weak and strong social network groups were defined by a median split of LSNS scores. Global cognitive functioning was evaluated by the Cognitive Abilities Screening Instrument (CASI), and depressive symptoms by the Center for Epidemiologic Studies Depression (CES-D) 11-item Scale. Dementia was diagnosed and classified by consensus of neurologists and geriatricians using standard criteria. Associations between baseline LSNS scores and the risks of DEM, AD and VaD were evaluated by multivariable Cox regression and Kaplan-Meier estimators using age as the time scale. Result: Study participants had a median (range) baseline age of 77 (71-93) years. The age-adjusted incidence of DEM was significantly higher among men who had weak (LSNS score ≤29) vs. strong (>29) social networks (12.6 vs. 8.7 per 1,000 person-years of follow-up; p = 0.014). Kaplan-Meier survival curves showed that the men with strong social networks were less likely to develop DEM (p<0.0001) and AD (p = 0.0006); the probability of DEM-free survival at almost 10 years was 93.8% for participants with strong social networks and 89.0% (11% incidence) for those with weak networks. In Cox regression models adjusting for baseline age, education, presence of at least one ApoE ɛ4 allele, prevalent stroke, depressive symptoms (CES-D score ≥9), and CASI score, weak social networks were significantly associated with increased risk of DEM (HR = 1.52, 95%CI = 1.11-2.08, p = 0.009) and AD (HR = 1.67, 95%CI = 1.11-2.51, p = 0.014). Conclusion: Strong social networks may protect against incident dementia and Alzheimer’s disease. Effective strategies should be devised to prevent the social isolation of older adults.