Vitality Watch

Dossier pubmed observational 11-may-2026
Human cohort, case-control, or epidemiological association

The impact of social isolation and loneliness on the excess risk for mortality, cardiovascular events, dementia, and depression among adults with stroke.

Key claim: In UK Biobank stroke survivors, social isolation and loneliness were independently associated with higher risks of all-cause and CVD mortality, dementia, and depression, with stronger associations than in people without stroke and partial mediation by health behaviours.

Abstract

BACKGROUND: Social isolation and loneliness, the rising social determinant of health, are prevalent in stroke survivors. We aimed to investigate whether social isolation and loneliness are associated with poor outcomes after stroke. METHODS: A total of 394,308 individuals (mean ± SD age, 55.9 ± 8.1 years, 55.7% female, 95.4% White) from the UK Biobank were included in this study. 7267 (1.8%) had a prevalent stroke at baseline, while the remaining 387,041 (98.2%) had no history of stroke. Stroke history was ascertained at baseline using both self-reported data and electronic health records. Social isolation and loneliness were assessed using self-reported questionnaires. Stroke history was defined by self-report and electronic health records. Post-stroke outcomes, including all-cause mortality, CVD mortality, all-cause dementia, CVD and depression, were identified by linking hospital records and death registries. RESULTS: During a median follow-up of 10.2 years, 2906 of the 7267 baseline stroke survivors experienced post-stroke outcomes. In fully adjusted models, social isolation was independently associated with higher risks of all-cause mortality (HR: 1.41, 95% CI: 1.20-1.64), CVD mortality (HR: 1.57, 1.26-1.96), and depression (HR: 1.35, 1.07-1.69). Loneliness was linked to increased risks of all-cause mortality (HR: 1.23, 1.01-1.50), CVD mortality (HR: 1.45, 1.11-1.87), all-cause dementia (HR: 1.67, 1.12-2.49), and depression (HR: 1.43, 1.09-1.88). These associations-particularly for all-cause mortality (Pinteraction = 0.006) and CVD mortality (Pinteraction = 0.042)-were more pronounced in stroke survivors than in those without stroke. Health behaviours mediated a substantial proportion of the associations for social isolation (32-55%) and loneliness (12-50%) with most outcomes. CONCLUSIONS: Social isolation and loneliness were associated with greater risks for developing premature death, CVD, dementia, and depression after stroke. The findings highlighted strengthening social connections to improve stroke prognosis.

Why this matters

Stroke survivors are an aging-dense population already carrying elevated risks for recurrent CVD, dementia, and depression, and this UK Biobank analysis shows the loneliness/isolation signal is not merely preserved but amplified relative to non-stroke adults, with health behaviours accounting for only part of the gap. That stronger-than-baseline association helps distinguish a true psychosocial exposure effect from the general background confounding that plagues social-connection epidemiology, while leaving reverse causation (post-stroke disability reshaping social life) unresolved.

Source

Link

Briefing card

The impact of social isolation and loneliness on the excess risk for mortality, cardiovascular events, dementia, and depression among adults with stroke.
Stage: observational · pubmed · 11-may-2026

If you or someone you care for is a stroke survivor, this cohort found social isolation and loneliness tracked with larger excess risks for death, dementia, and depression than the same exposures carry in people without stroke.

In UK Biobank stroke survivors, social isolation and loneliness were independently associated with higher risks of all-cause and CVD mortality, dementia, and depression, with stronger associations than in people without stroke and partial mediation by health behaviours.

Cross-references (0)

None recorded — doctrine links and citations appear here as scans and citation sweeps find them.

External: doi:10.1016/j.jad.2026.121949

build build 880 · 2bcef25 · 2026-11-18