Associations of loneliness with risk of Alzheimer's disease dementia in the Framingham Heart Study

Abstract

Introduction: The relationship between persistent loneliness and Alzheimer's disease (AD) is unclear. We examined the relationship between different types of mid-life loneliness and the development of dementia and AD.

Methods: Loneliness was assessed in cognitively normal adults using one item from the Center for Epidemiologic Studies Depression Scale. We defined loneliness as no loneliness, transient loneliness, incident loneliness,or persistent loneliness, and applied Cox regression models and Kaplan-Meier plots with dementia and AD as outcomes (n = 2880).

Results: After adjusting for demographics, social network, physical health, and apolipoprotein E ε4, persistent loneliness was associated with higher (hazard ratio [HR], 1.91; 95% confidence interval [CI] 1.25-2.90; P < .01), and transient loneliness with lower (HR, 0.34; 95% CI 0.14-0.84; P < .05), risk of dementia onset, compared to no loneliness. Results were similar for AD risk.

Discussion: Persistent loneliness in mid-life is an independent risk factor for dementia and AD, whereas recovery from loneliness suggests resilience to dementia risk.

Keywords: apolipoprotein E ε4; brain health; cognitive function; cohort study; depression; healthy aging; longitudinal; mental health; population-based; prevention; social isolation.

Figure 1:. Accumulative rates and Log-rank test analysis for dementia and AD based on loneliness status

Loneliness was defined as feeling lonely at least 1–2 days within the past week. No loneliness = participants did not report loneliness at neither exam 6 nor exam 7; Transient loneliness = participants reported loneliness only at exam 6; Incident loneliness = participants reported loneliness only at exam 7. Persistent loneliness = participants reported loneliness at both exam 6 and exam 7. The incident dementia (A) and the incident AD (B) between those with no loneliness and with different groups of loneliness were compared by using Chi-square (χ²) test. Kaplan-Meier method (C, D) was used to show the survival curves of up to 18 years of follow-up and the Log-rank test was used to compare the survival curves of loneliness with the no loneliness group as reference. Statistical significance is shown as * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001.

Figure 2.. Accumulative rates and Log-rank test analysis for AD based on loneliness status stratified by APOE ε4 genotype and depression status

Loneliness was defined as feeling lonely at least 1–2 days within the past week. No loneliness = participants did not report loneliness at neither exam 6 nor exam 7; Transient loneliness = participants reported loneliness only at exam 6; Persistent loneliness = participants reported loneliness at both exam 6 and exam 7. A-B and E-F: Subjects were stratified into those without (A) and with (B) APOE ε4 allele; those without (E) and with (F) depression (modified CES-D score ≥ 16). Then, they were further divided into those with no loneliness, with transient loneliness, and with persistent loneliness (X-axis). The incident rates of AD (Y-axis) between those with no loneliness and with different groups of loneliness in each subgroup were compared using Chi-square (χ²) test. C-D and G-H: Subjects were stratified into those without (C) and with (D) APOE ε4 allele; those without (G) and with (H) depression (modified CES-D score ≥ 16). They were further divided into no loneliness, transient loneliness, and persistent loneliness. Kaplan-Meier method was used to show the survival curves of up to 18 years of follow-up and the Log-rank test was used to compare the survival curves of loneliness with the no loneliness group as reference. Statistical significance is shown as * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001.