Abstract DP332: Loneliness is associated with 10-year risk of Stroke/Transient Ischemic Attack: The Framingham Heart Study

Introduction: Identifying modifiable targets is needed to reduce the increasing global burden of stroke. Psychosocial factors like loneliness are of interest given their rising prevalence worldwide, yet the temporal relationship of loneliness with subsequent stroke or transient ischemic attack (TIA) remains unclear. Hypothesis: Loneliness and changes in loneliness are associated with incident stroke/TIA. Methods: Framingham Heart Study (FHS) participants free of stroke/TIA, with two assessments of loneliness before incident stroke/TIA were included. Loneliness was assessed using “I felt lonely” item on Center for Epidemiologic Studies Depression Scale (CES-D) and defined as feeling lonely ≥3 days within the past week. We used two consecutive CES-D scores from different FHS exams to determine change in loneliness (mean interval of 6 years between exams). Loneliness change was defined as incident (loneliness at second exam), transient (loneliness at first exam), and persistent (loneliness at both exams). We collapsed these into two exposure-window categories based on recency: history of loneliness (transient&persistent) and current loneliness (incident&persistent). Because we observed age-specific interaction in association between loneliness and incident stroke/TIA, our analysis was stratified by age (<80, ≥80 years). Multivariable Cox regression models adjusted for age&sex, additionally vascular risk factors, depressive symptoms, and physical activity index, all at second exam. Sensitivity analysis explored association between loneliness change and stroke/TIA. Results: Among 4212 participants (mean age 62±10 years, 55% women), 138 (3%) had incident stroke/TIA over a median 5.4-year follow-up period (<80 years, 5.2-year; ≥80 years, 8.7-year). Overall, 174 (4.1%) people had history of loneliness, 217 (5.2%) had current loneliness. In all participants, those with a history of loneliness had >2-fold increase in stroke/TIA (p=0.02) which persisted after further adjustments. In those <80 years, a history of loneliness had >3-fold increase in stroke/TIA in all models. There was no association between current (more recent) loneliness and stroke/TIA. In sensitivity analysis, persistent (p=0.02) and transient loneliness (p<0.01) increased risk of stroke/TIA in those <80years, after adjustment. Conclusion: Our findings support the role for loneliness as an independent potentially modifiable risk factor for incident stroke/TIA in community dwelling individuals.