0736 Association of Excessive Sleepiness with All-cause Mortality Among Older Adult US Veterans

Abstract Introduction The association between excessive daytime sleepiness (EDS) and all-cause mortality remains uncertain. We investigated the relationship between EDS and all-cause mortality along with other factors including comorbid conditions. Methods We included patients who had visited the Veterans Health Administration (VHA) between FY 1999 and the end of 2022; and those who had been assigned ICD codes 9/10 of any sleep disorders or CPT codes of sleep studies. We developed and validated a Natural Language Processing (NLP) algorithm to extract Epworth sleepiness scale (ESS) from unstructured free-form text from physician notes. We assessed the NLP pipeline's performance using 470 patient notes that were manually annotated. We defined EDS as an ESS≥11. We categorized the patients into two groups: normal (ESS ≤10) and sleepy (ESS 10). Control factors were the most important variables determined through a machine learning algorithm. We reported adjusted odds ratios (aOR) and 95% confidence intervals (95%CI) from logistic regression models using most important factors. The process was separately applied to two distinct groups of veterans: those younger than 50 years and those aged 65 years or older. Results Out of 423,087 unique patients with ESS, 116,581 (27.6%) were older adults where 71,913 were grouped as normal and 44,668 as sleepy. Among young adults, only depression considered as the most important factor for all-cause mortality and all-cause mortality rates did not differ between normal and sleepy groups. For older adults, 33 factors including ESS were considered as the most important variables. The aOR of all-cause mortality was 21% higher in Sleepy compared Normal (aOR,1.21;95%CI:1.17,1.24). Among the comorbid conditions, the prevalence of diabetes, depression, heart failure, renal disease, PTSD, and cerebrovascular disease were higher in Sleepy compared to the Normal. Conclusion In sharp contrast to younger adults, excessive daytime sleepiness, as measured by ESS, predicted higher mortality in older adults. The presence of important comorbid conditions in older adults most likely plays a causal role. Chronic comorbid conditions may disturb sleep and result in excessive daytime sleepiness. Support (if any) Supported by NIH-NHLBI-K25 funding (#1K25HL152006-01 J.R.); seed funding by AHBPCE (#FY2024-002 J.R.); and Center for Innovations in Quality, Effectiveness, and Safety (CIN 13-413), MEDVAMC.