0695 Sleep And Neurocognitive Change In The Hispanic Community Health Study/Study Of Latinos (HCHS/SOL)

Abstract

Introduction To prospectively determine whether sleep-disordered breathing (SDB), sleep duration, insomnia and daytime sleepiness predict neurocognitive (NC) decline based on a representative sample of U.S. Hispanic/Latino adults. Methods We evaluated data from SOL-INCA (2016-2018) an ancillary to HCHS/SOL (Visit 1; 2008-2011). SOL-INCA (n=6,377) enrolled Hispanic/Latinos (45-74y) 5-7 years after baseline evaluation. The primary exposures were baseline measures of SDB (AHI≥15), insomnia assessed with the Women’s Health Initiative Insomnia Rating Scale score ≥10, self-reported sleep duration categorized into short ≤6hours (h), average >6-9h, long >9h and sleepiness with the Epworth sleepiness scale ≥10. NC outcomes measured change in episodic learning and memory (SEVLT-sum and SEVLT-Recall), language (WF; word fluency), processing speed (DSS; Digit Symbol Substitution), and a global cognitive impairment screener (SIS; Six-item Screener). We also considered cross-sectional associations with NC performance obtained in SOL-INCA. We used survey-based regression analyses to model sleep as a risk factor for NC performance and change. All NC measures were age-education adjusted z-scored. Models were further adjusted for sex, insurance, depression symptoms, vascular risk scores, study site and follow-up time. Results The mean age was 62±8 years, 54% were females; 7% were Central American, 25% were Cuban, 9.3% were Dominican, 35.6% were Mexican, and 14.1% were of Puerto Rican background. SDB was seen in 17% of the target population, while 6.4% had short sleep, 14.6% long sleep, 33.4% insomnia and 20% had sleepiness. Longer sleep, compared to >6-9h, was associated with worse NC outcomes at follow-up, and insomnia with worse memory. Compared to >6-9h, longer sleep was also associated with 6-year decline in global cognition (βSIS= -0.140 [SE=0.06]; p<0.05), episodic learning and memory (βSEVLT-Sum= -0.218 (se=0.06); p<0.001; βSEVLT-Recall= -0.123 (se=0.06); p<0.05) and verbal fluency (βWF= -0.182 [se=0.06]; p<0.05), but not processing speed. We found no associations between SDB, short sleep, sleepiness and cognitive performance at follow-up or change. Conclusion In a large, diverse community-based cohort of middle-aged and older U.S. Hispanic/Latinos, long sleep was related to worse cognitive performance and predicted neurocognitive decline over 6-years of follow-up. Support (If Any) AG056952, AG048642