0807 Associations Between Accelerometry-assessed Sleep Patterns and Hallmarks of Aging in the UK Biobank

Abstract Introduction Sleep duration, efficiency, and irregularity are associated with increased risks of cardiovascular, neurodegenerative, and immune-related diseases, highlighting sleep as a multidimensional construct. Sleep plays an important role in regulating diverse biological processes and shapes immune and genomic processes with relevance to biological aging. Evidence suggests that sleep disturbances may accelerate biological aging by triggering cellular and molecular impairments. However, few studies have examined how multidimensional sleep patterns linked to cellular and metabolic hallmarks of aging. Methods We analyzed accelerometer-derived sleep data from the UK Biobank and extracted six sleep characteristics using a validated self-supervised deep neural network: total sleep duration, REM, deep sleep, light sleep, wake after sleep onset (WASO), and sleep irregularity (SD of daily sleep duration). K-means clustering identified multidimensional sleep patterns. Diagnoses of 83 age-related diseases (ICD-10) were mapped to nine hallmarks of aging: genomic instability, telomere attrition, epigenetic alterations, loss of proteostasis, deregulated nutrient sensing, mitochondrial dysfunction, cellular senescence, stem cell exhaustion and altered intercellular communication. Cox proportional hazards models estimated associations between sleep patterns, age-related diseases and aging hallmarks. Results Among 93,249 participants, three distinct sleep patterns were identified, including pattern 1 distinguished by higher WASO and light sleep; pattern 2 marked by adequate total sleep duration and lower sleep irregularity; pattern 3 characterized by elevated sleep irregularity. Compared with pattern 2, pattern 1 was associated with increased risks of 18 diseases (P-FDR < 0.05), most notably Parkinson's disease (HR = 2.40, 95% CI = 1.82 to 3.05) and immunodeficiency (HR = 1.96, 95% CI = 1.00 to 3.82). Pattern 3 was associated with 24 diseases (P-FDR < 0.05), particularly Parkinson's disease (HR = 1.93, 95% CI = 1.48 to 2.53) and dementia (HR = 1.67, 95% CI = 1.38 to 2.02). Both cluster 1 and cluster 3 linked to higher risk of hallmarks of aging except genomic stability (P-FDR < 0.001). Conclusion Our findings identified three distinct sleep patterns using accelerometry-derived metrics in a large population cohort, each showing different associations with aging. These results demonstrated the importance of multidimensional sleep assessment and provide a framework for targeted strategies to promote healthy aging. Support (if any)